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}.
530 @section Single Stepping
532 @section Signal Handling
534 @section Thread Handling
536 @section Inferior Function Calls
538 @section Longjmp Support
540 @cindex @code{longjmp} debugging
541 @value{GDBN} has support for figuring out that the target is doing a
542 @code{longjmp} and for stopping at the target of the jump, if we are
543 stepping. This is done with a few specialized internal breakpoints,
544 which are visible in the output of the @samp{maint info breakpoint}
547 @findex GET_LONGJMP_TARGET
548 To make this work, you need to define a macro called
549 @code{GET_LONGJMP_TARGET}, which will examine the @code{jmp_buf}
550 structure and extract the longjmp target address. Since @code{jmp_buf}
551 is target specific, you will need to define it in the appropriate
552 @file{tm-@var{target}.h} file. Look in @file{tm-sun4os4.h} and
553 @file{sparc-tdep.c} for examples of how to do this.
558 Watchpoints are a special kind of breakpoints (@pxref{Algorithms,
559 breakpoints}) which break when data is accessed rather than when some
560 instruction is executed. When you have data which changes without
561 your knowing what code does that, watchpoints are the silver bullet to
562 hunt down and kill such bugs.
564 @cindex hardware watchpoints
565 @cindex software watchpoints
566 Watchpoints can be either hardware-assisted or not; the latter type is
567 known as ``software watchpoints.'' @value{GDBN} always uses
568 hardware-assisted watchpoints if they are available, and falls back on
569 software watchpoints otherwise. Typical situations where @value{GDBN}
570 will use software watchpoints are:
574 The watched memory region is too large for the underlying hardware
575 watchpoint support. For example, each x86 debug register can watch up
576 to 4 bytes of memory, so trying to watch data structures whose size is
577 more than 16 bytes will cause @value{GDBN} to use software
581 The value of the expression to be watched depends on data held in
582 registers (as opposed to memory).
585 Too many different watchpoints requested. (On some architectures,
586 this situation is impossible to detect until the debugged program is
587 resumed.) Note that x86 debug registers are used both for hardware
588 breakpoints and for watchpoints, so setting too many hardware
589 breakpoints might cause watchpoint insertion to fail.
592 No hardware-assisted watchpoints provided by the target
596 Software watchpoints are very slow, since @value{GDBN} needs to
597 single-step the program being debugged and test the value of the
598 watched expression(s) after each instruction. The rest of this
599 section is mostly irrelevant for software watchpoints.
601 When the inferior stops, @value{GDBN} tries to establish, among other
602 possible reasons, whether it stopped due to a watchpoint being hit.
603 For a data-write watchpoint, it does so by evaluating, for each
604 watchpoint, the expression whose value is being watched, and testing
605 whether the watched value has changed. For data-read and data-access
606 watchpoints, @value{GDBN} needs the target to supply a primitive that
607 returns the address of the data that was accessed or read (see the
608 description of @code{target_stopped_data_address} below): if this
609 primitive returns a valid address, @value{GDBN} infers that a
610 watchpoint triggered if it watches an expression whose evaluation uses
613 @value{GDBN} uses several macros and primitives to support hardware
617 @findex TARGET_HAS_HARDWARE_WATCHPOINTS
618 @item TARGET_HAS_HARDWARE_WATCHPOINTS
619 If defined, the target supports hardware watchpoints.
621 @findex TARGET_CAN_USE_HARDWARE_WATCHPOINT
622 @item TARGET_CAN_USE_HARDWARE_WATCHPOINT (@var{type}, @var{count}, @var{other})
623 Return the number of hardware watchpoints of type @var{type} that are
624 possible to be set. The value is positive if @var{count} watchpoints
625 of this type can be set, zero if setting watchpoints of this type is
626 not supported, and negative if @var{count} is more than the maximum
627 number of watchpoints of type @var{type} that can be set. @var{other}
628 is non-zero if other types of watchpoints are currently enabled (there
629 are architectures which cannot set watchpoints of different types at
632 @findex TARGET_REGION_OK_FOR_HW_WATCHPOINT
633 @item TARGET_REGION_OK_FOR_HW_WATCHPOINT (@var{addr}, @var{len})
634 Return non-zero if hardware watchpoints can be used to watch a region
635 whose address is @var{addr} and whose length in bytes is @var{len}.
637 @cindex insert or remove hardware watchpoint
638 @findex target_insert_watchpoint
639 @findex target_remove_watchpoint
640 @item target_insert_watchpoint (@var{addr}, @var{len}, @var{type})
641 @itemx target_remove_watchpoint (@var{addr}, @var{len}, @var{type})
642 Insert or remove a hardware watchpoint starting at @var{addr}, for
643 @var{len} bytes. @var{type} is the watchpoint type, one of the
644 possible values of the enumerated data type @code{target_hw_bp_type},
645 defined by @file{breakpoint.h} as follows:
648 enum target_hw_bp_type
650 hw_write = 0, /* Common (write) HW watchpoint */
651 hw_read = 1, /* Read HW watchpoint */
652 hw_access = 2, /* Access (read or write) HW watchpoint */
653 hw_execute = 3 /* Execute HW breakpoint */
658 These two macros should return 0 for success, non-zero for failure.
660 @cindex insert or remove hardware breakpoint
661 @findex target_remove_hw_breakpoint
662 @findex target_insert_hw_breakpoint
663 @item target_remove_hw_breakpoint (@var{addr}, @var{shadow})
664 @itemx target_insert_hw_breakpoint (@var{addr}, @var{shadow})
665 Insert or remove a hardware-assisted breakpoint at address @var{addr}.
666 Returns zero for success, non-zero for failure. @var{shadow} is the
667 real contents of the byte where the breakpoint has been inserted; it
668 is generally not valid when hardware breakpoints are used, but since
669 no other code touches these values, the implementations of the above
670 two macros can use them for their internal purposes.
672 @findex target_stopped_data_address
673 @item target_stopped_data_address (@var{addr_p})
674 If the inferior has some watchpoint that triggered, place the address
675 associated with the watchpoint at the location pointed to by
676 @var{addr_p} and return non-zero. Otherwise, return zero. Note that
677 this primitive is used by @value{GDBN} only on targets that support
678 data-read or data-access type watchpoints, so targets that have
679 support only for data-write watchpoints need not implement these
682 @findex HAVE_STEPPABLE_WATCHPOINT
683 @item HAVE_STEPPABLE_WATCHPOINT
684 If defined to a non-zero value, it is not necessary to disable a
685 watchpoint to step over it.
687 @findex HAVE_NONSTEPPABLE_WATCHPOINT
688 @item HAVE_NONSTEPPABLE_WATCHPOINT
689 If defined to a non-zero value, @value{GDBN} should disable a
690 watchpoint to step the inferior over it.
692 @findex HAVE_CONTINUABLE_WATCHPOINT
693 @item HAVE_CONTINUABLE_WATCHPOINT
694 If defined to a non-zero value, it is possible to continue the
695 inferior after a watchpoint has been hit.
697 @findex CANNOT_STEP_HW_WATCHPOINTS
698 @item CANNOT_STEP_HW_WATCHPOINTS
699 If this is defined to a non-zero value, @value{GDBN} will remove all
700 watchpoints before stepping the inferior.
702 @findex STOPPED_BY_WATCHPOINT
703 @item STOPPED_BY_WATCHPOINT (@var{wait_status})
704 Return non-zero if stopped by a watchpoint. @var{wait_status} is of
705 the type @code{struct target_waitstatus}, defined by @file{target.h}.
706 Normally, this macro is defined to invoke the function pointed to by
707 the @code{to_stopped_by_watchpoint} member of the structure (of the
708 type @code{target_ops}, defined on @file{target.h}) that describes the
709 target-specific operations; @code{to_stopped_by_watchpoint} ignores
710 the @var{wait_status} argument.
712 @value{GDBN} does not require the non-zero value returned by
713 @code{STOPPED_BY_WATCHPOINT} to be 100% correct, so if a target cannot
714 determine for sure whether the inferior stopped due to a watchpoint,
715 it could return non-zero ``just in case''.
718 @subsection x86 Watchpoints
719 @cindex x86 debug registers
720 @cindex watchpoints, on x86
722 The 32-bit Intel x86 (a.k.a.@: ia32) processors feature special debug
723 registers designed to facilitate debugging. @value{GDBN} provides a
724 generic library of functions that x86-based ports can use to implement
725 support for watchpoints and hardware-assisted breakpoints. This
726 subsection documents the x86 watchpoint facilities in @value{GDBN}.
728 To use the generic x86 watchpoint support, a port should do the
732 @findex I386_USE_GENERIC_WATCHPOINTS
734 Define the macro @code{I386_USE_GENERIC_WATCHPOINTS} somewhere in the
735 target-dependent headers.
738 Include the @file{config/i386/nm-i386.h} header file @emph{after}
739 defining @code{I386_USE_GENERIC_WATCHPOINTS}.
742 Add @file{i386-nat.o} to the value of the Make variable
743 @code{NATDEPFILES} (@pxref{Native Debugging, NATDEPFILES}) or
744 @code{TDEPFILES} (@pxref{Target Architecture Definition, TDEPFILES}).
747 Provide implementations for the @code{I386_DR_LOW_*} macros described
748 below. Typically, each macro should call a target-specific function
749 which does the real work.
752 The x86 watchpoint support works by maintaining mirror images of the
753 debug registers. Values are copied between the mirror images and the
754 real debug registers via a set of macros which each target needs to
758 @findex I386_DR_LOW_SET_CONTROL
759 @item I386_DR_LOW_SET_CONTROL (@var{val})
760 Set the Debug Control (DR7) register to the value @var{val}.
762 @findex I386_DR_LOW_SET_ADDR
763 @item I386_DR_LOW_SET_ADDR (@var{idx}, @var{addr})
764 Put the address @var{addr} into the debug register number @var{idx}.
766 @findex I386_DR_LOW_RESET_ADDR
767 @item I386_DR_LOW_RESET_ADDR (@var{idx})
768 Reset (i.e.@: zero out) the address stored in the debug register
771 @findex I386_DR_LOW_GET_STATUS
772 @item I386_DR_LOW_GET_STATUS
773 Return the value of the Debug Status (DR6) register. This value is
774 used immediately after it is returned by
775 @code{I386_DR_LOW_GET_STATUS}, so as to support per-thread status
779 For each one of the 4 debug registers (whose indices are from 0 to 3)
780 that store addresses, a reference count is maintained by @value{GDBN},
781 to allow sharing of debug registers by several watchpoints. This
782 allows users to define several watchpoints that watch the same
783 expression, but with different conditions and/or commands, without
784 wasting debug registers which are in short supply. @value{GDBN}
785 maintains the reference counts internally, targets don't have to do
786 anything to use this feature.
788 The x86 debug registers can each watch a region that is 1, 2, or 4
789 bytes long. The ia32 architecture requires that each watched region
790 be appropriately aligned: 2-byte region on 2-byte boundary, 4-byte
791 region on 4-byte boundary. However, the x86 watchpoint support in
792 @value{GDBN} can watch unaligned regions and regions larger than 4
793 bytes (up to 16 bytes) by allocating several debug registers to watch
794 a single region. This allocation of several registers per a watched
795 region is also done automatically without target code intervention.
797 The generic x86 watchpoint support provides the following API for the
798 @value{GDBN}'s application code:
801 @findex i386_region_ok_for_watchpoint
802 @item i386_region_ok_for_watchpoint (@var{addr}, @var{len})
803 The macro @code{TARGET_REGION_OK_FOR_HW_WATCHPOINT} is set to call
804 this function. It counts the number of debug registers required to
805 watch a given region, and returns a non-zero value if that number is
806 less than 4, the number of debug registers available to x86
809 @findex i386_stopped_data_address
810 @item i386_stopped_data_address (@var{addr_p})
812 @code{target_stopped_data_address} is set to call this function.
814 function examines the breakpoint condition bits in the DR6 Debug
815 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
816 macro, and returns the address associated with the first bit that is
819 @findex i386_stopped_by_watchpoint
820 @item i386_stopped_by_watchpoint (void)
821 The macro @code{STOPPED_BY_WATCHPOINT}
822 is set to call this function. The
823 argument passed to @code{STOPPED_BY_WATCHPOINT} is ignored. This
824 function examines the breakpoint condition bits in the DR6 Debug
825 Status register, as returned by the @code{I386_DR_LOW_GET_STATUS}
826 macro, and returns true if any bit is set. Otherwise, false is
829 @findex i386_insert_watchpoint
830 @findex i386_remove_watchpoint
831 @item i386_insert_watchpoint (@var{addr}, @var{len}, @var{type})
832 @itemx i386_remove_watchpoint (@var{addr}, @var{len}, @var{type})
833 Insert or remove a watchpoint. The macros
834 @code{target_insert_watchpoint} and @code{target_remove_watchpoint}
835 are set to call these functions. @code{i386_insert_watchpoint} first
836 looks for a debug register which is already set to watch the same
837 region for the same access types; if found, it just increments the
838 reference count of that debug register, thus implementing debug
839 register sharing between watchpoints. If no such register is found,
840 the function looks for a vacant debug register, sets its mirrored
841 value to @var{addr}, sets the mirrored value of DR7 Debug Control
842 register as appropriate for the @var{len} and @var{type} parameters,
843 and then passes the new values of the debug register and DR7 to the
844 inferior by calling @code{I386_DR_LOW_SET_ADDR} and
845 @code{I386_DR_LOW_SET_CONTROL}. If more than one debug register is
846 required to cover the given region, the above process is repeated for
849 @code{i386_remove_watchpoint} does the opposite: it resets the address
850 in the mirrored value of the debug register and its read/write and
851 length bits in the mirrored value of DR7, then passes these new
852 values to the inferior via @code{I386_DR_LOW_RESET_ADDR} and
853 @code{I386_DR_LOW_SET_CONTROL}. If a register is shared by several
854 watchpoints, each time a @code{i386_remove_watchpoint} is called, it
855 decrements the reference count, and only calls
856 @code{I386_DR_LOW_RESET_ADDR} and @code{I386_DR_LOW_SET_CONTROL} when
857 the count goes to zero.
859 @findex i386_insert_hw_breakpoint
860 @findex i386_remove_hw_breakpoint
861 @item i386_insert_hw_breakpoint (@var{addr}, @var{shadow}
862 @itemx i386_remove_hw_breakpoint (@var{addr}, @var{shadow})
863 These functions insert and remove hardware-assisted breakpoints. The
864 macros @code{target_insert_hw_breakpoint} and
865 @code{target_remove_hw_breakpoint} are set to call these functions.
866 These functions work like @code{i386_insert_watchpoint} and
867 @code{i386_remove_watchpoint}, respectively, except that they set up
868 the debug registers to watch instruction execution, and each
869 hardware-assisted breakpoint always requires exactly one debug
872 @findex i386_stopped_by_hwbp
873 @item i386_stopped_by_hwbp (void)
874 This function returns non-zero if the inferior has some watchpoint or
875 hardware breakpoint that triggered. It works like
876 @code{i386_stopped_data_address}, except that it doesn't record the
877 address whose watchpoint triggered.
879 @findex i386_cleanup_dregs
880 @item i386_cleanup_dregs (void)
881 This function clears all the reference counts, addresses, and control
882 bits in the mirror images of the debug registers. It doesn't affect
883 the actual debug registers in the inferior process.
890 x86 processors support setting watchpoints on I/O reads or writes.
891 However, since no target supports this (as of March 2001), and since
892 @code{enum target_hw_bp_type} doesn't even have an enumeration for I/O
893 watchpoints, this feature is not yet available to @value{GDBN} running
897 x86 processors can enable watchpoints locally, for the current task
898 only, or globally, for all the tasks. For each debug register,
899 there's a bit in the DR7 Debug Control register that determines
900 whether the associated address is watched locally or globally. The
901 current implementation of x86 watchpoint support in @value{GDBN}
902 always sets watchpoints to be locally enabled, since global
903 watchpoints might interfere with the underlying OS and are probably
904 unavailable in many platforms.
910 In the abstract, a checkpoint is a point in the execution history of
911 the program, which the user may wish to return to at some later time.
913 Internally, a checkpoint is a saved copy of the program state, including
914 whatever information is required in order to restore the program to that
915 state at a later time. This can be expected to include the state of
916 registers and memory, and may include external state such as the state
917 of open files and devices.
919 There are a number of ways in which checkpoints may be implemented
920 in gdb, eg. as corefiles, as forked processes, and as some opaque
921 method implemented on the target side.
923 A corefile can be used to save an image of target memory and register
924 state, which can in principle be restored later --- but corefiles do
925 not typically include information about external entities such as
926 open files. Currently this method is not implemented in gdb.
928 A forked process can save the state of user memory and registers,
929 as well as some subset of external (kernel) state. This method
930 is used to implement checkpoints on Linux, and in principle might
931 be used on other systems.
933 Some targets, eg.@: simulators, might have their own built-in
934 method for saving checkpoints, and gdb might be able to take
935 advantage of that capability without necessarily knowing any
936 details of how it is done.
939 @section Observing changes in @value{GDBN} internals
940 @cindex observer pattern interface
941 @cindex notifications about changes in internals
943 In order to function properly, several modules need to be notified when
944 some changes occur in the @value{GDBN} internals. Traditionally, these
945 modules have relied on several paradigms, the most common ones being
946 hooks and gdb-events. Unfortunately, none of these paradigms was
947 versatile enough to become the standard notification mechanism in
948 @value{GDBN}. The fact that they only supported one ``client'' was also
951 A new paradigm, based on the Observer pattern of the @cite{Design
952 Patterns} book, has therefore been implemented. The goal was to provide
953 a new interface overcoming the issues with the notification mechanisms
954 previously available. This new interface needed to be strongly typed,
955 easy to extend, and versatile enough to be used as the standard
956 interface when adding new notifications.
958 See @ref{GDB Observers} for a brief description of the observers
959 currently implemented in GDB. The rationale for the current
960 implementation is also briefly discussed.
964 @chapter User Interface
966 @value{GDBN} has several user interfaces. Although the command-line interface
967 is the most common and most familiar, there are others.
969 @section Command Interpreter
971 @cindex command interpreter
973 The command interpreter in @value{GDBN} is fairly simple. It is designed to
974 allow for the set of commands to be augmented dynamically, and also
975 has a recursive subcommand capability, where the first argument to
976 a command may itself direct a lookup on a different command list.
978 For instance, the @samp{set} command just starts a lookup on the
979 @code{setlist} command list, while @samp{set thread} recurses
980 to the @code{set_thread_cmd_list}.
984 To add commands in general, use @code{add_cmd}. @code{add_com} adds to
985 the main command list, and should be used for those commands. The usual
986 place to add commands is in the @code{_initialize_@var{xyz}} routines at
987 the ends of most source files.
989 @findex add_setshow_cmd
990 @findex add_setshow_cmd_full
991 To add paired @samp{set} and @samp{show} commands, use
992 @code{add_setshow_cmd} or @code{add_setshow_cmd_full}. The former is
993 a slightly simpler interface which is useful when you don't need to
994 further modify the new command structures, while the latter returns
995 the new command structures for manipulation.
997 @cindex deprecating commands
998 @findex deprecate_cmd
999 Before removing commands from the command set it is a good idea to
1000 deprecate them for some time. Use @code{deprecate_cmd} on commands or
1001 aliases to set the deprecated flag. @code{deprecate_cmd} takes a
1002 @code{struct cmd_list_element} as it's first argument. You can use the
1003 return value from @code{add_com} or @code{add_cmd} to deprecate the
1004 command immediately after it is created.
1006 The first time a command is used the user will be warned and offered a
1007 replacement (if one exists). Note that the replacement string passed to
1008 @code{deprecate_cmd} should be the full name of the command, i.e. the
1009 entire string the user should type at the command line.
1011 @section UI-Independent Output---the @code{ui_out} Functions
1012 @c This section is based on the documentation written by Fernando
1013 @c Nasser <fnasser@redhat.com>.
1015 @cindex @code{ui_out} functions
1016 The @code{ui_out} functions present an abstraction level for the
1017 @value{GDBN} output code. They hide the specifics of different user
1018 interfaces supported by @value{GDBN}, and thus free the programmer
1019 from the need to write several versions of the same code, one each for
1020 every UI, to produce output.
1022 @subsection Overview and Terminology
1024 In general, execution of each @value{GDBN} command produces some sort
1025 of output, and can even generate an input request.
1027 Output can be generated for the following purposes:
1031 to display a @emph{result} of an operation;
1034 to convey @emph{info} or produce side-effects of a requested
1038 to provide a @emph{notification} of an asynchronous event (including
1039 progress indication of a prolonged asynchronous operation);
1042 to display @emph{error messages} (including warnings);
1045 to show @emph{debug data};
1048 to @emph{query} or prompt a user for input (a special case).
1052 This section mainly concentrates on how to build result output,
1053 although some of it also applies to other kinds of output.
1055 Generation of output that displays the results of an operation
1056 involves one or more of the following:
1060 output of the actual data
1063 formatting the output as appropriate for console output, to make it
1064 easily readable by humans
1067 machine oriented formatting--a more terse formatting to allow for easy
1068 parsing by programs which read @value{GDBN}'s output
1071 annotation, whose purpose is to help legacy GUIs to identify interesting
1075 The @code{ui_out} routines take care of the first three aspects.
1076 Annotations are provided by separate annotation routines. Note that use
1077 of annotations for an interface between a GUI and @value{GDBN} is
1080 Output can be in the form of a single item, which we call a @dfn{field};
1081 a @dfn{list} consisting of identical fields; a @dfn{tuple} consisting of
1082 non-identical fields; or a @dfn{table}, which is a tuple consisting of a
1083 header and a body. In a BNF-like form:
1086 @item <table> @expansion{}
1087 @code{<header> <body>}
1088 @item <header> @expansion{}
1089 @code{@{ <column> @}}
1090 @item <column> @expansion{}
1091 @code{<width> <alignment> <title>}
1092 @item <body> @expansion{}
1097 @subsection General Conventions
1099 Most @code{ui_out} routines are of type @code{void}, the exceptions are
1100 @code{ui_out_stream_new} (which returns a pointer to the newly created
1101 object) and the @code{make_cleanup} routines.
1103 The first parameter is always the @code{ui_out} vector object, a pointer
1104 to a @code{struct ui_out}.
1106 The @var{format} parameter is like in @code{printf} family of functions.
1107 When it is present, there must also be a variable list of arguments
1108 sufficient used to satisfy the @code{%} specifiers in the supplied
1111 When a character string argument is not used in a @code{ui_out} function
1112 call, a @code{NULL} pointer has to be supplied instead.
1115 @subsection Table, Tuple and List Functions
1117 @cindex list output functions
1118 @cindex table output functions
1119 @cindex tuple output functions
1120 This section introduces @code{ui_out} routines for building lists,
1121 tuples and tables. The routines to output the actual data items
1122 (fields) are presented in the next section.
1124 To recap: A @dfn{tuple} is a sequence of @dfn{fields}, each field
1125 containing information about an object; a @dfn{list} is a sequence of
1126 fields where each field describes an identical object.
1128 Use the @dfn{table} functions when your output consists of a list of
1129 rows (tuples) and the console output should include a heading. Use this
1130 even when you are listing just one object but you still want the header.
1132 @cindex nesting level in @code{ui_out} functions
1133 Tables can not be nested. Tuples and lists can be nested up to a
1134 maximum of five levels.
1136 The overall structure of the table output code is something like this:
1151 Here is the description of table-, tuple- and list-related @code{ui_out}
1154 @deftypefun void ui_out_table_begin (struct ui_out *@var{uiout}, int @var{nbrofcols}, int @var{nr_rows}, const char *@var{tblid})
1155 The function @code{ui_out_table_begin} marks the beginning of the output
1156 of a table. It should always be called before any other @code{ui_out}
1157 function for a given table. @var{nbrofcols} is the number of columns in
1158 the table. @var{nr_rows} is the number of rows in the table.
1159 @var{tblid} is an optional string identifying the table. The string
1160 pointed to by @var{tblid} is copied by the implementation of
1161 @code{ui_out_table_begin}, so the application can free the string if it
1162 was @code{malloc}ed.
1164 The companion function @code{ui_out_table_end}, described below, marks
1165 the end of the table's output.
1168 @deftypefun void ui_out_table_header (struct ui_out *@var{uiout}, int @var{width}, enum ui_align @var{alignment}, const char *@var{colhdr})
1169 @code{ui_out_table_header} provides the header information for a single
1170 table column. You call this function several times, one each for every
1171 column of the table, after @code{ui_out_table_begin}, but before
1172 @code{ui_out_table_body}.
1174 The value of @var{width} gives the column width in characters. The
1175 value of @var{alignment} is one of @code{left}, @code{center}, and
1176 @code{right}, and it specifies how to align the header: left-justify,
1177 center, or right-justify it. @var{colhdr} points to a string that
1178 specifies the column header; the implementation copies that string, so
1179 column header strings in @code{malloc}ed storage can be freed after the
1183 @deftypefun void ui_out_table_body (struct ui_out *@var{uiout})
1184 This function delimits the table header from the table body.
1187 @deftypefun void ui_out_table_end (struct ui_out *@var{uiout})
1188 This function signals the end of a table's output. It should be called
1189 after the table body has been produced by the list and field output
1192 There should be exactly one call to @code{ui_out_table_end} for each
1193 call to @code{ui_out_table_begin}, otherwise the @code{ui_out} functions
1194 will signal an internal error.
1197 The output of the tuples that represent the table rows must follow the
1198 call to @code{ui_out_table_body} and precede the call to
1199 @code{ui_out_table_end}. You build a tuple by calling
1200 @code{ui_out_tuple_begin} and @code{ui_out_tuple_end}, with suitable
1201 calls to functions which actually output fields between them.
1203 @deftypefun void ui_out_tuple_begin (struct ui_out *@var{uiout}, const char *@var{id})
1204 This function marks the beginning of a tuple output. @var{id} points
1205 to an optional string that identifies the tuple; it is copied by the
1206 implementation, and so strings in @code{malloc}ed storage can be freed
1210 @deftypefun void ui_out_tuple_end (struct ui_out *@var{uiout})
1211 This function signals an end of a tuple output. There should be exactly
1212 one call to @code{ui_out_tuple_end} for each call to
1213 @code{ui_out_tuple_begin}, otherwise an internal @value{GDBN} error will
1217 @deftypefun struct cleanup *make_cleanup_ui_out_tuple_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1218 This function first opens the tuple and then establishes a cleanup
1219 (@pxref{Coding, Cleanups}) to close the tuple. It provides a convenient
1220 and correct implementation of the non-portable@footnote{The function
1221 cast is not portable ISO C.} code sequence:
1223 struct cleanup *old_cleanup;
1224 ui_out_tuple_begin (uiout, "...");
1225 old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
1230 @deftypefun void ui_out_list_begin (struct ui_out *@var{uiout}, const char *@var{id})
1231 This function marks the beginning of a list output. @var{id} points to
1232 an optional string that identifies the list; it is copied by the
1233 implementation, and so strings in @code{malloc}ed storage can be freed
1237 @deftypefun void ui_out_list_end (struct ui_out *@var{uiout})
1238 This function signals an end of a list output. There should be exactly
1239 one call to @code{ui_out_list_end} for each call to
1240 @code{ui_out_list_begin}, otherwise an internal @value{GDBN} error will
1244 @deftypefun struct cleanup *make_cleanup_ui_out_list_begin_end (struct ui_out *@var{uiout}, const char *@var{id})
1245 Similar to @code{make_cleanup_ui_out_tuple_begin_end}, this function
1246 opens a list and then establishes cleanup (@pxref{Coding, Cleanups})
1247 that will close the list.list.
1250 @subsection Item Output Functions
1252 @cindex item output functions
1253 @cindex field output functions
1255 The functions described below produce output for the actual data
1256 items, or fields, which contain information about the object.
1258 Choose the appropriate function accordingly to your particular needs.
1260 @deftypefun void ui_out_field_fmt (struct ui_out *@var{uiout}, char *@var{fldname}, char *@var{format}, ...)
1261 This is the most general output function. It produces the
1262 representation of the data in the variable-length argument list
1263 according to formatting specifications in @var{format}, a
1264 @code{printf}-like format string. The optional argument @var{fldname}
1265 supplies the name of the field. The data items themselves are
1266 supplied as additional arguments after @var{format}.
1268 This generic function should be used only when it is not possible to
1269 use one of the specialized versions (see below).
1272 @deftypefun void ui_out_field_int (struct ui_out *@var{uiout}, const char *@var{fldname}, int @var{value})
1273 This function outputs a value of an @code{int} variable. It uses the
1274 @code{"%d"} output conversion specification. @var{fldname} specifies
1275 the name of the field.
1278 @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})
1279 This function outputs a value of an @code{int} variable. It differs from
1280 @code{ui_out_field_int} in that the caller specifies the desired @var{width} and @var{alignment} of the output.
1281 @var{fldname} specifies
1282 the name of the field.
1285 @deftypefun void ui_out_field_core_addr (struct ui_out *@var{uiout}, const char *@var{fldname}, CORE_ADDR @var{address})
1286 This function outputs an address.
1289 @deftypefun void ui_out_field_string (struct ui_out *@var{uiout}, const char *@var{fldname}, const char *@var{string})
1290 This function outputs a string using the @code{"%s"} conversion
1294 Sometimes, there's a need to compose your output piece by piece using
1295 functions that operate on a stream, such as @code{value_print} or
1296 @code{fprintf_symbol_filtered}. These functions accept an argument of
1297 the type @code{struct ui_file *}, a pointer to a @code{ui_file} object
1298 used to store the data stream used for the output. When you use one
1299 of these functions, you need a way to pass their results stored in a
1300 @code{ui_file} object to the @code{ui_out} functions. To this end,
1301 you first create a @code{ui_stream} object by calling
1302 @code{ui_out_stream_new}, pass the @code{stream} member of that
1303 @code{ui_stream} object to @code{value_print} and similar functions,
1304 and finally call @code{ui_out_field_stream} to output the field you
1305 constructed. When the @code{ui_stream} object is no longer needed,
1306 you should destroy it and free its memory by calling
1307 @code{ui_out_stream_delete}.
1309 @deftypefun struct ui_stream *ui_out_stream_new (struct ui_out *@var{uiout})
1310 This function creates a new @code{ui_stream} object which uses the
1311 same output methods as the @code{ui_out} object whose pointer is
1312 passed in @var{uiout}. It returns a pointer to the newly created
1313 @code{ui_stream} object.
1316 @deftypefun void ui_out_stream_delete (struct ui_stream *@var{streambuf})
1317 This functions destroys a @code{ui_stream} object specified by
1321 @deftypefun void ui_out_field_stream (struct ui_out *@var{uiout}, const char *@var{fieldname}, struct ui_stream *@var{streambuf})
1322 This function consumes all the data accumulated in
1323 @code{streambuf->stream} and outputs it like
1324 @code{ui_out_field_string} does. After a call to
1325 @code{ui_out_field_stream}, the accumulated data no longer exists, but
1326 the stream is still valid and may be used for producing more fields.
1329 @strong{Important:} If there is any chance that your code could bail
1330 out before completing output generation and reaching the point where
1331 @code{ui_out_stream_delete} is called, it is necessary to set up a
1332 cleanup, to avoid leaking memory and other resources. Here's a
1333 skeleton code to do that:
1336 struct ui_stream *mybuf = ui_out_stream_new (uiout);
1337 struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
1342 If the function already has the old cleanup chain set (for other kinds
1343 of cleanups), you just have to add your cleanup to it:
1346 mybuf = ui_out_stream_new (uiout);
1347 make_cleanup (ui_out_stream_delete, mybuf);
1350 Note that with cleanups in place, you should not call
1351 @code{ui_out_stream_delete} directly, or you would attempt to free the
1354 @subsection Utility Output Functions
1356 @deftypefun void ui_out_field_skip (struct ui_out *@var{uiout}, const char *@var{fldname})
1357 This function skips a field in a table. Use it if you have to leave
1358 an empty field without disrupting the table alignment. The argument
1359 @var{fldname} specifies a name for the (missing) filed.
1362 @deftypefun void ui_out_text (struct ui_out *@var{uiout}, const char *@var{string})
1363 This function outputs the text in @var{string} in a way that makes it
1364 easy to be read by humans. For example, the console implementation of
1365 this method filters the text through a built-in pager, to prevent it
1366 from scrolling off the visible portion of the screen.
1368 Use this function for printing relatively long chunks of text around
1369 the actual field data: the text it produces is not aligned according
1370 to the table's format. Use @code{ui_out_field_string} to output a
1371 string field, and use @code{ui_out_message}, described below, to
1372 output short messages.
1375 @deftypefun void ui_out_spaces (struct ui_out *@var{uiout}, int @var{nspaces})
1376 This function outputs @var{nspaces} spaces. It is handy to align the
1377 text produced by @code{ui_out_text} with the rest of the table or
1381 @deftypefun void ui_out_message (struct ui_out *@var{uiout}, int @var{verbosity}, const char *@var{format}, ...)
1382 This function produces a formatted message, provided that the current
1383 verbosity level is at least as large as given by @var{verbosity}. The
1384 current verbosity level is specified by the user with the @samp{set
1385 verbositylevel} command.@footnote{As of this writing (April 2001),
1386 setting verbosity level is not yet implemented, and is always returned
1387 as zero. So calling @code{ui_out_message} with a @var{verbosity}
1388 argument more than zero will cause the message to never be printed.}
1391 @deftypefun void ui_out_wrap_hint (struct ui_out *@var{uiout}, char *@var{indent})
1392 This function gives the console output filter (a paging filter) a hint
1393 of where to break lines which are too long. Ignored for all other
1394 output consumers. @var{indent}, if non-@code{NULL}, is the string to
1395 be printed to indent the wrapped text on the next line; it must remain
1396 accessible until the next call to @code{ui_out_wrap_hint}, or until an
1397 explicit newline is produced by one of the other functions. If
1398 @var{indent} is @code{NULL}, the wrapped text will not be indented.
1401 @deftypefun void ui_out_flush (struct ui_out *@var{uiout})
1402 This function flushes whatever output has been accumulated so far, if
1403 the UI buffers output.
1407 @subsection Examples of Use of @code{ui_out} functions
1409 @cindex using @code{ui_out} functions
1410 @cindex @code{ui_out} functions, usage examples
1411 This section gives some practical examples of using the @code{ui_out}
1412 functions to generalize the old console-oriented code in
1413 @value{GDBN}. The examples all come from functions defined on the
1414 @file{breakpoints.c} file.
1416 This example, from the @code{breakpoint_1} function, shows how to
1419 The original code was:
1422 if (!found_a_breakpoint++)
1424 annotate_breakpoints_headers ();
1427 printf_filtered ("Num ");
1429 printf_filtered ("Type ");
1431 printf_filtered ("Disp ");
1433 printf_filtered ("Enb ");
1437 printf_filtered ("Address ");
1440 printf_filtered ("What\n");
1442 annotate_breakpoints_table ();
1446 Here's the new version:
1449 nr_printable_breakpoints = @dots{};
1452 ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
1454 ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
1456 if (nr_printable_breakpoints > 0)
1457 annotate_breakpoints_headers ();
1458 if (nr_printable_breakpoints > 0)
1460 ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
1461 if (nr_printable_breakpoints > 0)
1463 ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
1464 if (nr_printable_breakpoints > 0)
1466 ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
1467 if (nr_printable_breakpoints > 0)
1469 ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
1472 if (nr_printable_breakpoints > 0)
1474 if (TARGET_ADDR_BIT <= 32)
1475 ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
1477 ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
1479 if (nr_printable_breakpoints > 0)
1481 ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
1482 ui_out_table_body (uiout);
1483 if (nr_printable_breakpoints > 0)
1484 annotate_breakpoints_table ();
1487 This example, from the @code{print_one_breakpoint} function, shows how
1488 to produce the actual data for the table whose structure was defined
1489 in the above example. The original code was:
1494 printf_filtered ("%-3d ", b->number);
1496 if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
1497 || ((int) b->type != bptypes[(int) b->type].type))
1498 internal_error ("bptypes table does not describe type #%d.",
1500 printf_filtered ("%-14s ", bptypes[(int)b->type].description);
1502 printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
1504 printf_filtered ("%-3c ", bpenables[(int)b->enable]);
1508 This is the new version:
1512 ui_out_tuple_begin (uiout, "bkpt");
1514 ui_out_field_int (uiout, "number", b->number);
1516 if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
1517 || ((int) b->type != bptypes[(int) b->type].type))
1518 internal_error ("bptypes table does not describe type #%d.",
1520 ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
1522 ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
1524 ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
1528 This example, also from @code{print_one_breakpoint}, shows how to
1529 produce a complicated output field using the @code{print_expression}
1530 functions which requires a stream to be passed. It also shows how to
1531 automate stream destruction with cleanups. The original code was:
1535 print_expression (b->exp, gdb_stdout);
1541 struct ui_stream *stb = ui_out_stream_new (uiout);
1542 struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
1545 print_expression (b->exp, stb->stream);
1546 ui_out_field_stream (uiout, "what", local_stream);
1549 This example, also from @code{print_one_breakpoint}, shows how to use
1550 @code{ui_out_text} and @code{ui_out_field_string}. The original code
1555 if (b->dll_pathname == NULL)
1556 printf_filtered ("<any library> ");
1558 printf_filtered ("library \"%s\" ", b->dll_pathname);
1565 if (b->dll_pathname == NULL)
1567 ui_out_field_string (uiout, "what", "<any library>");
1568 ui_out_spaces (uiout, 1);
1572 ui_out_text (uiout, "library \"");
1573 ui_out_field_string (uiout, "what", b->dll_pathname);
1574 ui_out_text (uiout, "\" ");
1578 The following example from @code{print_one_breakpoint} shows how to
1579 use @code{ui_out_field_int} and @code{ui_out_spaces}. The original
1584 if (b->forked_inferior_pid != 0)
1585 printf_filtered ("process %d ", b->forked_inferior_pid);
1592 if (b->forked_inferior_pid != 0)
1594 ui_out_text (uiout, "process ");
1595 ui_out_field_int (uiout, "what", b->forked_inferior_pid);
1596 ui_out_spaces (uiout, 1);
1600 Here's an example of using @code{ui_out_field_string}. The original
1605 if (b->exec_pathname != NULL)
1606 printf_filtered ("program \"%s\" ", b->exec_pathname);
1613 if (b->exec_pathname != NULL)
1615 ui_out_text (uiout, "program \"");
1616 ui_out_field_string (uiout, "what", b->exec_pathname);
1617 ui_out_text (uiout, "\" ");
1621 Finally, here's an example of printing an address. The original code:
1625 printf_filtered ("%s ",
1626 hex_string_custom ((unsigned long) b->address, 8));
1633 ui_out_field_core_addr (uiout, "Address", b->address);
1637 @section Console Printing
1646 @cindex @code{libgdb}
1647 @code{libgdb} 1.0 was an abortive project of years ago. The theory was
1648 to provide an API to @value{GDBN}'s functionality.
1651 @cindex @code{libgdb}
1652 @code{libgdb} 2.0 is an ongoing effort to update @value{GDBN} so that is
1653 better able to support graphical and other environments.
1655 Since @code{libgdb} development is on-going, its architecture is still
1656 evolving. The following components have so far been identified:
1660 Observer - @file{gdb-events.h}.
1662 Builder - @file{ui-out.h}
1664 Event Loop - @file{event-loop.h}
1666 Library - @file{gdb.h}
1669 The model that ties these components together is described below.
1671 @section The @code{libgdb} Model
1673 A client of @code{libgdb} interacts with the library in two ways.
1677 As an observer (using @file{gdb-events}) receiving notifications from
1678 @code{libgdb} of any internal state changes (break point changes, run
1681 As a client querying @code{libgdb} (using the @file{ui-out} builder) to
1682 obtain various status values from @value{GDBN}.
1685 Since @code{libgdb} could have multiple clients (e.g., a GUI supporting
1686 the existing @value{GDBN} CLI), those clients must co-operate when
1687 controlling @code{libgdb}. In particular, a client must ensure that
1688 @code{libgdb} is idle (i.e. no other client is using @code{libgdb})
1689 before responding to a @file{gdb-event} by making a query.
1691 @section CLI support
1693 At present @value{GDBN}'s CLI is very much entangled in with the core of
1694 @code{libgdb}. Consequently, a client wishing to include the CLI in
1695 their interface needs to carefully co-ordinate its own and the CLI's
1698 It is suggested that the client set @code{libgdb} up to be bi-modal
1699 (alternate between CLI and client query modes). The notes below sketch
1704 The client registers itself as an observer of @code{libgdb}.
1706 The client create and install @code{cli-out} builder using its own
1707 versions of the @code{ui-file} @code{gdb_stderr}, @code{gdb_stdtarg} and
1708 @code{gdb_stdout} streams.
1710 The client creates a separate custom @code{ui-out} builder that is only
1711 used while making direct queries to @code{libgdb}.
1714 When the client receives input intended for the CLI, it simply passes it
1715 along. Since the @code{cli-out} builder is installed by default, all
1716 the CLI output in response to that command is routed (pronounced rooted)
1717 through to the client controlled @code{gdb_stdout} et.@: al.@: streams.
1718 At the same time, the client is kept abreast of internal changes by
1719 virtue of being a @code{libgdb} observer.
1721 The only restriction on the client is that it must wait until
1722 @code{libgdb} becomes idle before initiating any queries (using the
1723 client's custom builder).
1725 @section @code{libgdb} components
1727 @subheading Observer - @file{gdb-events.h}
1728 @file{gdb-events} provides the client with a very raw mechanism that can
1729 be used to implement an observer. At present it only allows for one
1730 observer and that observer must, internally, handle the need to delay
1731 the processing of any event notifications until after @code{libgdb} has
1732 finished the current command.
1734 @subheading Builder - @file{ui-out.h}
1735 @file{ui-out} provides the infrastructure necessary for a client to
1736 create a builder. That builder is then passed down to @code{libgdb}
1737 when doing any queries.
1739 @subheading Event Loop - @file{event-loop.h}
1740 @c There could be an entire section on the event-loop
1741 @file{event-loop}, currently non-re-entrant, provides a simple event
1742 loop. A client would need to either plug its self into this loop or,
1743 implement a new event-loop that GDB would use.
1745 The event-loop will eventually be made re-entrant. This is so that
1746 @value{GDBN} can better handle the problem of some commands blocking
1747 instead of returning.
1749 @subheading Library - @file{gdb.h}
1750 @file{libgdb} is the most obvious component of this system. It provides
1751 the query interface. Each function is parameterized by a @code{ui-out}
1752 builder. The result of the query is constructed using that builder
1753 before the query function returns.
1755 @node Symbol Handling
1757 @chapter Symbol Handling
1759 Symbols are a key part of @value{GDBN}'s operation. Symbols include variables,
1760 functions, and types.
1762 @section Symbol Reading
1764 @cindex symbol reading
1765 @cindex reading of symbols
1766 @cindex symbol files
1767 @value{GDBN} reads symbols from @dfn{symbol files}. The usual symbol
1768 file is the file containing the program which @value{GDBN} is
1769 debugging. @value{GDBN} can be directed to use a different file for
1770 symbols (with the @samp{symbol-file} command), and it can also read
1771 more symbols via the @samp{add-file} and @samp{load} commands, or while
1772 reading symbols from shared libraries.
1774 @findex find_sym_fns
1775 Symbol files are initially opened by code in @file{symfile.c} using
1776 the BFD library (@pxref{Support Libraries}). BFD identifies the type
1777 of the file by examining its header. @code{find_sym_fns} then uses
1778 this identification to locate a set of symbol-reading functions.
1780 @findex add_symtab_fns
1781 @cindex @code{sym_fns} structure
1782 @cindex adding a symbol-reading module
1783 Symbol-reading modules identify themselves to @value{GDBN} by calling
1784 @code{add_symtab_fns} during their module initialization. The argument
1785 to @code{add_symtab_fns} is a @code{struct sym_fns} which contains the
1786 name (or name prefix) of the symbol format, the length of the prefix,
1787 and pointers to four functions. These functions are called at various
1788 times to process symbol files whose identification matches the specified
1791 The functions supplied by each module are:
1794 @item @var{xyz}_symfile_init(struct sym_fns *sf)
1796 @cindex secondary symbol file
1797 Called from @code{symbol_file_add} when we are about to read a new
1798 symbol file. This function should clean up any internal state (possibly
1799 resulting from half-read previous files, for example) and prepare to
1800 read a new symbol file. Note that the symbol file which we are reading
1801 might be a new ``main'' symbol file, or might be a secondary symbol file
1802 whose symbols are being added to the existing symbol table.
1804 The argument to @code{@var{xyz}_symfile_init} is a newly allocated
1805 @code{struct sym_fns} whose @code{bfd} field contains the BFD for the
1806 new symbol file being read. Its @code{private} field has been zeroed,
1807 and can be modified as desired. Typically, a struct of private
1808 information will be @code{malloc}'d, and a pointer to it will be placed
1809 in the @code{private} field.
1811 There is no result from @code{@var{xyz}_symfile_init}, but it can call
1812 @code{error} if it detects an unavoidable problem.
1814 @item @var{xyz}_new_init()
1816 Called from @code{symbol_file_add} when discarding existing symbols.
1817 This function needs only handle the symbol-reading module's internal
1818 state; the symbol table data structures visible to the rest of
1819 @value{GDBN} will be discarded by @code{symbol_file_add}. It has no
1820 arguments and no result. It may be called after
1821 @code{@var{xyz}_symfile_init}, if a new symbol table is being read, or
1822 may be called alone if all symbols are simply being discarded.
1824 @item @var{xyz}_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)
1826 Called from @code{symbol_file_add} to actually read the symbols from a
1827 symbol-file into a set of psymtabs or symtabs.
1829 @code{sf} points to the @code{struct sym_fns} originally passed to
1830 @code{@var{xyz}_sym_init} for possible initialization. @code{addr} is
1831 the offset between the file's specified start address and its true
1832 address in memory. @code{mainline} is 1 if this is the main symbol
1833 table being read, and 0 if a secondary symbol file (e.g., shared library
1834 or dynamically loaded file) is being read.@refill
1837 In addition, if a symbol-reading module creates psymtabs when
1838 @var{xyz}_symfile_read is called, these psymtabs will contain a pointer
1839 to a function @code{@var{xyz}_psymtab_to_symtab}, which can be called
1840 from any point in the @value{GDBN} symbol-handling code.
1843 @item @var{xyz}_psymtab_to_symtab (struct partial_symtab *pst)
1845 Called from @code{psymtab_to_symtab} (or the @code{PSYMTAB_TO_SYMTAB} macro) if
1846 the psymtab has not already been read in and had its @code{pst->symtab}
1847 pointer set. The argument is the psymtab to be fleshed-out into a
1848 symtab. Upon return, @code{pst->readin} should have been set to 1, and
1849 @code{pst->symtab} should contain a pointer to the new corresponding symtab, or
1850 zero if there were no symbols in that part of the symbol file.
1853 @section Partial Symbol Tables
1855 @value{GDBN} has three types of symbol tables:
1858 @cindex full symbol table
1861 Full symbol tables (@dfn{symtabs}). These contain the main
1862 information about symbols and addresses.
1866 Partial symbol tables (@dfn{psymtabs}). These contain enough
1867 information to know when to read the corresponding part of the full
1870 @cindex minimal symbol table
1873 Minimal symbol tables (@dfn{msymtabs}). These contain information
1874 gleaned from non-debugging symbols.
1877 @cindex partial symbol table
1878 This section describes partial symbol tables.
1880 A psymtab is constructed by doing a very quick pass over an executable
1881 file's debugging information. Small amounts of information are
1882 extracted---enough to identify which parts of the symbol table will
1883 need to be re-read and fully digested later, when the user needs the
1884 information. The speed of this pass causes @value{GDBN} to start up very
1885 quickly. Later, as the detailed rereading occurs, it occurs in small
1886 pieces, at various times, and the delay therefrom is mostly invisible to
1888 @c (@xref{Symbol Reading}.)
1890 The symbols that show up in a file's psymtab should be, roughly, those
1891 visible to the debugger's user when the program is not running code from
1892 that file. These include external symbols and types, static symbols and
1893 types, and @code{enum} values declared at file scope.
1895 The psymtab also contains the range of instruction addresses that the
1896 full symbol table would represent.
1898 @cindex finding a symbol
1899 @cindex symbol lookup
1900 The idea is that there are only two ways for the user (or much of the
1901 code in the debugger) to reference a symbol:
1904 @findex find_pc_function
1905 @findex find_pc_line
1907 By its address (e.g., execution stops at some address which is inside a
1908 function in this file). The address will be noticed to be in the
1909 range of this psymtab, and the full symtab will be read in.
1910 @code{find_pc_function}, @code{find_pc_line}, and other
1911 @code{find_pc_@dots{}} functions handle this.
1913 @cindex lookup_symbol
1916 (e.g., the user asks to print a variable, or set a breakpoint on a
1917 function). Global names and file-scope names will be found in the
1918 psymtab, which will cause the symtab to be pulled in. Local names will
1919 have to be qualified by a global name, or a file-scope name, in which
1920 case we will have already read in the symtab as we evaluated the
1921 qualifier. Or, a local symbol can be referenced when we are ``in'' a
1922 local scope, in which case the first case applies. @code{lookup_symbol}
1923 does most of the work here.
1926 The only reason that psymtabs exist is to cause a symtab to be read in
1927 at the right moment. Any symbol that can be elided from a psymtab,
1928 while still causing that to happen, should not appear in it. Since
1929 psymtabs don't have the idea of scope, you can't put local symbols in
1930 them anyway. Psymtabs don't have the idea of the type of a symbol,
1931 either, so types need not appear, unless they will be referenced by
1934 It is a bug for @value{GDBN} to behave one way when only a psymtab has
1935 been read, and another way if the corresponding symtab has been read
1936 in. Such bugs are typically caused by a psymtab that does not contain
1937 all the visible symbols, or which has the wrong instruction address
1940 The psymtab for a particular section of a symbol file (objfile) could be
1941 thrown away after the symtab has been read in. The symtab should always
1942 be searched before the psymtab, so the psymtab will never be used (in a
1943 bug-free environment). Currently, psymtabs are allocated on an obstack,
1944 and all the psymbols themselves are allocated in a pair of large arrays
1945 on an obstack, so there is little to be gained by trying to free them
1946 unless you want to do a lot more work.
1950 @unnumberedsubsec Fundamental Types (e.g., @code{FT_VOID}, @code{FT_BOOLEAN}).
1952 @cindex fundamental types
1953 These are the fundamental types that @value{GDBN} uses internally. Fundamental
1954 types from the various debugging formats (stabs, ELF, etc) are mapped
1955 into one of these. They are basically a union of all fundamental types
1956 that @value{GDBN} knows about for all the languages that @value{GDBN}
1959 @unnumberedsubsec Type Codes (e.g., @code{TYPE_CODE_PTR}, @code{TYPE_CODE_ARRAY}).
1962 Each time @value{GDBN} builds an internal type, it marks it with one
1963 of these types. The type may be a fundamental type, such as
1964 @code{TYPE_CODE_INT}, or a derived type, such as @code{TYPE_CODE_PTR}
1965 which is a pointer to another type. Typically, several @code{FT_*}
1966 types map to one @code{TYPE_CODE_*} type, and are distinguished by
1967 other members of the type struct, such as whether the type is signed
1968 or unsigned, and how many bits it uses.
1970 @unnumberedsubsec Builtin Types (e.g., @code{builtin_type_void}, @code{builtin_type_char}).
1972 These are instances of type structs that roughly correspond to
1973 fundamental types and are created as global types for @value{GDBN} to
1974 use for various ugly historical reasons. We eventually want to
1975 eliminate these. Note for example that @code{builtin_type_int}
1976 initialized in @file{gdbtypes.c} is basically the same as a
1977 @code{TYPE_CODE_INT} type that is initialized in @file{c-lang.c} for
1978 an @code{FT_INTEGER} fundamental type. The difference is that the
1979 @code{builtin_type} is not associated with any particular objfile, and
1980 only one instance exists, while @file{c-lang.c} builds as many
1981 @code{TYPE_CODE_INT} types as needed, with each one associated with
1982 some particular objfile.
1984 @section Object File Formats
1985 @cindex object file formats
1989 @cindex @code{a.out} format
1990 The @code{a.out} format is the original file format for Unix. It
1991 consists of three sections: @code{text}, @code{data}, and @code{bss},
1992 which are for program code, initialized data, and uninitialized data,
1995 The @code{a.out} format is so simple that it doesn't have any reserved
1996 place for debugging information. (Hey, the original Unix hackers used
1997 @samp{adb}, which is a machine-language debugger!) The only debugging
1998 format for @code{a.out} is stabs, which is encoded as a set of normal
1999 symbols with distinctive attributes.
2001 The basic @code{a.out} reader is in @file{dbxread.c}.
2006 The COFF format was introduced with System V Release 3 (SVR3) Unix.
2007 COFF files may have multiple sections, each prefixed by a header. The
2008 number of sections is limited.
2010 The COFF specification includes support for debugging. Although this
2011 was a step forward, the debugging information was woefully limited. For
2012 instance, it was not possible to represent code that came from an
2015 The COFF reader is in @file{coffread.c}.
2019 @cindex ECOFF format
2020 ECOFF is an extended COFF originally introduced for Mips and Alpha
2023 The basic ECOFF reader is in @file{mipsread.c}.
2027 @cindex XCOFF format
2028 The IBM RS/6000 running AIX uses an object file format called XCOFF.
2029 The COFF sections, symbols, and line numbers are used, but debugging
2030 symbols are @code{dbx}-style stabs whose strings are located in the
2031 @code{.debug} section (rather than the string table). For more
2032 information, see @ref{Top,,,stabs,The Stabs Debugging Format}.
2034 The shared library scheme has a clean interface for figuring out what
2035 shared libraries are in use, but the catch is that everything which
2036 refers to addresses (symbol tables and breakpoints at least) needs to be
2037 relocated for both shared libraries and the main executable. At least
2038 using the standard mechanism this can only be done once the program has
2039 been run (or the core file has been read).
2043 @cindex PE-COFF format
2044 Windows 95 and NT use the PE (@dfn{Portable Executable}) format for their
2045 executables. PE is basically COFF with additional headers.
2047 While BFD includes special PE support, @value{GDBN} needs only the basic
2053 The ELF format came with System V Release 4 (SVR4) Unix. ELF is similar
2054 to COFF in being organized into a number of sections, but it removes
2055 many of COFF's limitations.
2057 The basic ELF reader is in @file{elfread.c}.
2062 SOM is HP's object file and debug format (not to be confused with IBM's
2063 SOM, which is a cross-language ABI).
2065 The SOM reader is in @file{hpread.c}.
2067 @subsection Other File Formats
2069 @cindex Netware Loadable Module format
2070 Other file formats that have been supported by @value{GDBN} include Netware
2071 Loadable Modules (@file{nlmread.c}).
2073 @section Debugging File Formats
2075 This section describes characteristics of debugging information that
2076 are independent of the object file format.
2080 @cindex stabs debugging info
2081 @code{stabs} started out as special symbols within the @code{a.out}
2082 format. Since then, it has been encapsulated into other file
2083 formats, such as COFF and ELF.
2085 While @file{dbxread.c} does some of the basic stab processing,
2086 including for encapsulated versions, @file{stabsread.c} does
2091 @cindex COFF debugging info
2092 The basic COFF definition includes debugging information. The level
2093 of support is minimal and non-extensible, and is not often used.
2095 @subsection Mips debug (Third Eye)
2097 @cindex ECOFF debugging info
2098 ECOFF includes a definition of a special debug format.
2100 The file @file{mdebugread.c} implements reading for this format.
2104 @cindex DWARF 1 debugging info
2105 DWARF 1 is a debugging format that was originally designed to be
2106 used with ELF in SVR4 systems.
2111 @c If defined, these are the producer strings in a DWARF 1 file. All of
2112 @c these have reasonable defaults already.
2114 The DWARF 1 reader is in @file{dwarfread.c}.
2118 @cindex DWARF 2 debugging info
2119 DWARF 2 is an improved but incompatible version of DWARF 1.
2121 The DWARF 2 reader is in @file{dwarf2read.c}.
2125 @cindex SOM debugging info
2126 Like COFF, the SOM definition includes debugging information.
2128 @section Adding a New Symbol Reader to @value{GDBN}
2130 @cindex adding debugging info reader
2131 If you are using an existing object file format (@code{a.out}, COFF, ELF, etc),
2132 there is probably little to be done.
2134 If you need to add a new object file format, you must first add it to
2135 BFD. This is beyond the scope of this document.
2137 You must then arrange for the BFD code to provide access to the
2138 debugging symbols. Generally @value{GDBN} will have to call swapping routines
2139 from BFD and a few other BFD internal routines to locate the debugging
2140 information. As much as possible, @value{GDBN} should not depend on the BFD
2141 internal data structures.
2143 For some targets (e.g., COFF), there is a special transfer vector used
2144 to call swapping routines, since the external data structures on various
2145 platforms have different sizes and layouts. Specialized routines that
2146 will only ever be implemented by one object file format may be called
2147 directly. This interface should be described in a file
2148 @file{bfd/lib@var{xyz}.h}, which is included by @value{GDBN}.
2150 @section Memory Management for Symbol Files
2152 Most memory associated with a loaded symbol file is stored on
2153 its @code{objfile_obstack}. This includes symbols, types,
2154 namespace data, and other information produced by the symbol readers.
2156 Because this data lives on the objfile's obstack, it is automatically
2157 released when the objfile is unloaded or reloaded. Therefore one
2158 objfile must not reference symbol or type data from another objfile;
2159 they could be unloaded at different times.
2161 User convenience variables, et cetera, have associated types. Normally
2162 these types live in the associated objfile. However, when the objfile
2163 is unloaded, those types are deep copied to global memory, so that
2164 the values of the user variables and history items are not lost.
2167 @node Language Support
2169 @chapter Language Support
2171 @cindex language support
2172 @value{GDBN}'s language support is mainly driven by the symbol reader,
2173 although it is possible for the user to set the source language
2176 @value{GDBN} chooses the source language by looking at the extension
2177 of the file recorded in the debug info; @file{.c} means C, @file{.f}
2178 means Fortran, etc. It may also use a special-purpose language
2179 identifier if the debug format supports it, like with DWARF.
2181 @section Adding a Source Language to @value{GDBN}
2183 @cindex adding source language
2184 To add other languages to @value{GDBN}'s expression parser, follow the
2188 @item Create the expression parser.
2190 @cindex expression parser
2191 This should reside in a file @file{@var{lang}-exp.y}. Routines for
2192 building parsed expressions into a @code{union exp_element} list are in
2195 @cindex language parser
2196 Since we can't depend upon everyone having Bison, and YACC produces
2197 parsers that define a bunch of global names, the following lines
2198 @strong{must} be included at the top of the YACC parser, to prevent the
2199 various parsers from defining the same global names:
2202 #define yyparse @var{lang}_parse
2203 #define yylex @var{lang}_lex
2204 #define yyerror @var{lang}_error
2205 #define yylval @var{lang}_lval
2206 #define yychar @var{lang}_char
2207 #define yydebug @var{lang}_debug
2208 #define yypact @var{lang}_pact
2209 #define yyr1 @var{lang}_r1
2210 #define yyr2 @var{lang}_r2
2211 #define yydef @var{lang}_def
2212 #define yychk @var{lang}_chk
2213 #define yypgo @var{lang}_pgo
2214 #define yyact @var{lang}_act
2215 #define yyexca @var{lang}_exca
2216 #define yyerrflag @var{lang}_errflag
2217 #define yynerrs @var{lang}_nerrs
2220 At the bottom of your parser, define a @code{struct language_defn} and
2221 initialize it with the right values for your language. Define an
2222 @code{initialize_@var{lang}} routine and have it call
2223 @samp{add_language(@var{lang}_language_defn)} to tell the rest of @value{GDBN}
2224 that your language exists. You'll need some other supporting variables
2225 and functions, which will be used via pointers from your
2226 @code{@var{lang}_language_defn}. See the declaration of @code{struct
2227 language_defn} in @file{language.h}, and the other @file{*-exp.y} files,
2228 for more information.
2230 @item Add any evaluation routines, if necessary
2232 @cindex expression evaluation routines
2233 @findex evaluate_subexp
2234 @findex prefixify_subexp
2235 @findex length_of_subexp
2236 If you need new opcodes (that represent the operations of the language),
2237 add them to the enumerated type in @file{expression.h}. Add support
2238 code for these operations in the @code{evaluate_subexp} function
2239 defined in the file @file{eval.c}. Add cases
2240 for new opcodes in two functions from @file{parse.c}:
2241 @code{prefixify_subexp} and @code{length_of_subexp}. These compute
2242 the number of @code{exp_element}s that a given operation takes up.
2244 @item Update some existing code
2246 Add an enumerated identifier for your language to the enumerated type
2247 @code{enum language} in @file{defs.h}.
2249 Update the routines in @file{language.c} so your language is included.
2250 These routines include type predicates and such, which (in some cases)
2251 are language dependent. If your language does not appear in the switch
2252 statement, an error is reported.
2254 @vindex current_language
2255 Also included in @file{language.c} is the code that updates the variable
2256 @code{current_language}, and the routines that translate the
2257 @code{language_@var{lang}} enumerated identifier into a printable
2260 @findex _initialize_language
2261 Update the function @code{_initialize_language} to include your
2262 language. This function picks the default language upon startup, so is
2263 dependent upon which languages that @value{GDBN} is built for.
2265 @findex allocate_symtab
2266 Update @code{allocate_symtab} in @file{symfile.c} and/or symbol-reading
2267 code so that the language of each symtab (source file) is set properly.
2268 This is used to determine the language to use at each stack frame level.
2269 Currently, the language is set based upon the extension of the source
2270 file. If the language can be better inferred from the symbol
2271 information, please set the language of the symtab in the symbol-reading
2274 @findex print_subexp
2275 @findex op_print_tab
2276 Add helper code to @code{print_subexp} (in @file{expprint.c}) to handle any new
2277 expression opcodes you have added to @file{expression.h}. Also, add the
2278 printed representations of your operators to @code{op_print_tab}.
2280 @item Add a place of call
2283 Add a call to @code{@var{lang}_parse()} and @code{@var{lang}_error} in
2284 @code{parse_exp_1} (defined in @file{parse.c}).
2286 @item Use macros to trim code
2288 @cindex trimming language-dependent code
2289 The user has the option of building @value{GDBN} for some or all of the
2290 languages. If the user decides to build @value{GDBN} for the language
2291 @var{lang}, then every file dependent on @file{language.h} will have the
2292 macro @code{_LANG_@var{lang}} defined in it. Use @code{#ifdef}s to
2293 leave out large routines that the user won't need if he or she is not
2294 using your language.
2296 Note that you do not need to do this in your YACC parser, since if @value{GDBN}
2297 is not build for @var{lang}, then @file{@var{lang}-exp.tab.o} (the
2298 compiled form of your parser) is not linked into @value{GDBN} at all.
2300 See the file @file{configure.in} for how @value{GDBN} is configured
2301 for different languages.
2303 @item Edit @file{Makefile.in}
2305 Add dependencies in @file{Makefile.in}. Make sure you update the macro
2306 variables such as @code{HFILES} and @code{OBJS}, otherwise your code may
2307 not get linked in, or, worse yet, it may not get @code{tar}red into the
2312 @node Host Definition
2314 @chapter Host Definition
2316 With the advent of Autoconf, it's rarely necessary to have host
2317 definition machinery anymore. The following information is provided,
2318 mainly, as an historical reference.
2320 @section Adding a New Host
2322 @cindex adding a new host
2323 @cindex host, adding
2324 @value{GDBN}'s host configuration support normally happens via Autoconf.
2325 New host-specific definitions should not be needed. Older hosts
2326 @value{GDBN} still use the host-specific definitions and files listed
2327 below, but these mostly exist for historical reasons, and will
2328 eventually disappear.
2331 @item gdb/config/@var{arch}/@var{xyz}.mh
2332 This file once contained both host and native configuration information
2333 (@pxref{Native Debugging}) for the machine @var{xyz}. The host
2334 configuration information is now handed by Autoconf.
2336 Host configuration information included a definition of
2337 @code{XM_FILE=xm-@var{xyz}.h} and possibly definitions for @code{CC},
2338 @code{SYSV_DEFINE}, @code{XM_CFLAGS}, @code{XM_ADD_FILES},
2339 @code{XM_CLIBS}, @code{XM_CDEPS}, etc.; see @file{Makefile.in}.
2341 New host only configurations do not need this file.
2343 @item gdb/config/@var{arch}/xm-@var{xyz}.h
2344 This file once contained definitions and includes required when hosting
2345 gdb on machine @var{xyz}. Those definitions and includes are now
2346 handled by Autoconf.
2348 New host and native configurations do not need this file.
2350 @emph{Maintainer's note: Some hosts continue to use the @file{xm-xyz.h}
2351 file to define the macros @var{HOST_FLOAT_FORMAT},
2352 @var{HOST_DOUBLE_FORMAT} and @var{HOST_LONG_DOUBLE_FORMAT}. That code
2353 also needs to be replaced with either an Autoconf or run-time test.}
2357 @subheading Generic Host Support Files
2359 @cindex generic host support
2360 There are some ``generic'' versions of routines that can be used by
2361 various systems. These can be customized in various ways by macros
2362 defined in your @file{xm-@var{xyz}.h} file. If these routines work for
2363 the @var{xyz} host, you can just include the generic file's name (with
2364 @samp{.o}, not @samp{.c}) in @code{XDEPFILES}.
2366 Otherwise, if your machine needs custom support routines, you will need
2367 to write routines that perform the same functions as the generic file.
2368 Put them into @code{@var{xyz}-xdep.c}, and put @code{@var{xyz}-xdep.o}
2369 into @code{XDEPFILES}.
2372 @cindex remote debugging support
2373 @cindex serial line support
2375 This contains serial line support for Unix systems. This is always
2376 included, via the makefile variable @code{SER_HARDWIRE}; override this
2377 variable in the @file{.mh} file to avoid it.
2380 This contains serial line support for 32-bit programs running under DOS,
2381 using the DJGPP (a.k.a.@: GO32) execution environment.
2383 @cindex TCP remote support
2385 This contains generic TCP support using sockets.
2388 @section Host Conditionals
2390 When @value{GDBN} is configured and compiled, various macros are
2391 defined or left undefined, to control compilation based on the
2392 attributes of the host system. These macros and their meanings (or if
2393 the meaning is not documented here, then one of the source files where
2394 they are used is indicated) are:
2397 @item @value{GDBN}INIT_FILENAME
2398 The default name of @value{GDBN}'s initialization file (normally
2402 This macro is deprecated.
2404 @item SIGWINCH_HANDLER
2405 If your host defines @code{SIGWINCH}, you can define this to be the name
2406 of a function to be called if @code{SIGWINCH} is received.
2408 @item SIGWINCH_HANDLER_BODY
2409 Define this to expand into code that will define the function named by
2410 the expansion of @code{SIGWINCH_HANDLER}.
2412 @item ALIGN_STACK_ON_STARTUP
2413 @cindex stack alignment
2414 Define this if your system is of a sort that will crash in
2415 @code{tgetent} if the stack happens not to be longword-aligned when
2416 @code{main} is called. This is a rare situation, but is known to occur
2417 on several different types of systems.
2419 @item CRLF_SOURCE_FILES
2420 @cindex DOS text files
2421 Define this if host files use @code{\r\n} rather than @code{\n} as a
2422 line terminator. This will cause source file listings to omit @code{\r}
2423 characters when printing and it will allow @code{\r\n} line endings of files
2424 which are ``sourced'' by gdb. It must be possible to open files in binary
2425 mode using @code{O_BINARY} or, for fopen, @code{"rb"}.
2427 @item DEFAULT_PROMPT
2429 The default value of the prompt string (normally @code{"(gdb) "}).
2432 @cindex terminal device
2433 The name of the generic TTY device, defaults to @code{"/dev/tty"}.
2436 Define this if binary files are opened the same way as text files.
2440 In some cases, use the system call @code{mmap} for reading symbol
2441 tables. For some machines this allows for sharing and quick updates.
2444 Define this if the host system has @code{termio.h}.
2451 Values for host-side constants.
2454 Substitute for isatty, if not available.
2457 This is the longest integer type available on the host. If not defined,
2458 it will default to @code{long long} or @code{long}, depending on
2459 @code{CC_HAS_LONG_LONG}.
2461 @item CC_HAS_LONG_LONG
2462 @cindex @code{long long} data type
2463 Define this if the host C compiler supports @code{long long}. This is set
2464 by the @code{configure} script.
2466 @item PRINTF_HAS_LONG_LONG
2467 Define this if the host can handle printing of long long integers via
2468 the printf format conversion specifier @code{ll}. This is set by the
2469 @code{configure} script.
2471 @item HAVE_LONG_DOUBLE
2472 Define this if the host C compiler supports @code{long double}. This is
2473 set by the @code{configure} script.
2475 @item PRINTF_HAS_LONG_DOUBLE
2476 Define this if the host can handle printing of long double float-point
2477 numbers via the printf format conversion specifier @code{Lg}. This is
2478 set by the @code{configure} script.
2480 @item SCANF_HAS_LONG_DOUBLE
2481 Define this if the host can handle the parsing of long double
2482 float-point numbers via the scanf format conversion specifier
2483 @code{Lg}. This is set by the @code{configure} script.
2485 @item LSEEK_NOT_LINEAR
2486 Define this if @code{lseek (n)} does not necessarily move to byte number
2487 @code{n} in the file. This is only used when reading source files. It
2488 is normally faster to define @code{CRLF_SOURCE_FILES} when possible.
2491 This macro is used as the argument to @code{lseek} (or, most commonly,
2492 @code{bfd_seek}). FIXME, should be replaced by SEEK_SET instead,
2493 which is the POSIX equivalent.
2496 If defined, this should be one or more tokens, such as @code{volatile},
2497 that can be used in both the declaration and definition of functions to
2498 indicate that they never return. The default is already set correctly
2499 if compiling with GCC. This will almost never need to be defined.
2502 If defined, this should be one or more tokens, such as
2503 @code{__attribute__ ((noreturn))}, that can be used in the declarations
2504 of functions to indicate that they never return. The default is already
2505 set correctly if compiling with GCC. This will almost never need to be
2510 Define these to appropriate value for the system @code{lseek}, if not already
2514 This is the signal for stopping @value{GDBN}. Defaults to
2515 @code{SIGTSTP}. (Only redefined for the Convex.)
2518 Means that System V (prior to SVR4) include files are in use. (FIXME:
2519 This symbol is abused in @file{infrun.c}, @file{regex.c}, and
2520 @file{utils.c} for other things, at the moment.)
2523 Define this to help placate @code{lint} in some situations.
2526 Define this to override the defaults of @code{__volatile__} or
2531 @node Target Architecture Definition
2533 @chapter Target Architecture Definition
2535 @cindex target architecture definition
2536 @value{GDBN}'s target architecture defines what sort of
2537 machine-language programs @value{GDBN} can work with, and how it works
2540 The target architecture object is implemented as the C structure
2541 @code{struct gdbarch *}. The structure, and its methods, are generated
2542 using the Bourne shell script @file{gdbarch.sh}.
2544 @section Operating System ABI Variant Handling
2545 @cindex OS ABI variants
2547 @value{GDBN} provides a mechanism for handling variations in OS
2548 ABIs. An OS ABI variant may have influence over any number of
2549 variables in the target architecture definition. There are two major
2550 components in the OS ABI mechanism: sniffers and handlers.
2552 A @dfn{sniffer} examines a file matching a BFD architecture/flavour pair
2553 (the architecture may be wildcarded) in an attempt to determine the
2554 OS ABI of that file. Sniffers with a wildcarded architecture are considered
2555 to be @dfn{generic}, while sniffers for a specific architecture are
2556 considered to be @dfn{specific}. A match from a specific sniffer
2557 overrides a match from a generic sniffer. Multiple sniffers for an
2558 architecture/flavour may exist, in order to differentiate between two
2559 different operating systems which use the same basic file format. The
2560 OS ABI framework provides a generic sniffer for ELF-format files which
2561 examines the @code{EI_OSABI} field of the ELF header, as well as note
2562 sections known to be used by several operating systems.
2564 @cindex fine-tuning @code{gdbarch} structure
2565 A @dfn{handler} is used to fine-tune the @code{gdbarch} structure for the
2566 selected OS ABI. There may be only one handler for a given OS ABI
2567 for each BFD architecture.
2569 The following OS ABI variants are defined in @file{osabi.h}:
2573 @findex GDB_OSABI_UNKNOWN
2574 @item GDB_OSABI_UNKNOWN
2575 The ABI of the inferior is unknown. The default @code{gdbarch}
2576 settings for the architecture will be used.
2578 @findex GDB_OSABI_SVR4
2579 @item GDB_OSABI_SVR4
2580 UNIX System V Release 4
2582 @findex GDB_OSABI_HURD
2583 @item GDB_OSABI_HURD
2584 GNU using the Hurd kernel
2586 @findex GDB_OSABI_SOLARIS
2587 @item GDB_OSABI_SOLARIS
2590 @findex GDB_OSABI_OSF1
2591 @item GDB_OSABI_OSF1
2592 OSF/1, including Digital UNIX and Compaq Tru64 UNIX
2594 @findex GDB_OSABI_LINUX
2595 @item GDB_OSABI_LINUX
2596 GNU using the Linux kernel
2598 @findex GDB_OSABI_FREEBSD_AOUT
2599 @item GDB_OSABI_FREEBSD_AOUT
2600 FreeBSD using the a.out executable format
2602 @findex GDB_OSABI_FREEBSD_ELF
2603 @item GDB_OSABI_FREEBSD_ELF
2604 FreeBSD using the ELF executable format
2606 @findex GDB_OSABI_NETBSD_AOUT
2607 @item GDB_OSABI_NETBSD_AOUT
2608 NetBSD using the a.out executable format
2610 @findex GDB_OSABI_NETBSD_ELF
2611 @item GDB_OSABI_NETBSD_ELF
2612 NetBSD using the ELF executable format
2614 @findex GDB_OSABI_WINCE
2615 @item GDB_OSABI_WINCE
2618 @findex GDB_OSABI_GO32
2619 @item GDB_OSABI_GO32
2622 @findex GDB_OSABI_NETWARE
2623 @item GDB_OSABI_NETWARE
2626 @findex GDB_OSABI_ARM_EABI_V1
2627 @item GDB_OSABI_ARM_EABI_V1
2628 ARM Embedded ABI version 1
2630 @findex GDB_OSABI_ARM_EABI_V2
2631 @item GDB_OSABI_ARM_EABI_V2
2632 ARM Embedded ABI version 2
2634 @findex GDB_OSABI_ARM_APCS
2635 @item GDB_OSABI_ARM_APCS
2636 Generic ARM Procedure Call Standard
2640 Here are the functions that make up the OS ABI framework:
2642 @deftypefun const char *gdbarch_osabi_name (enum gdb_osabi @var{osabi})
2643 Return the name of the OS ABI corresponding to @var{osabi}.
2646 @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}))
2647 Register the OS ABI handler specified by @var{init_osabi} for the
2648 architecture, machine type and OS ABI specified by @var{arch},
2649 @var{machine} and @var{osabi}. In most cases, a value of zero for the
2650 machine type, which implies the architecture's default machine type,
2654 @deftypefun void gdbarch_register_osabi_sniffer (enum bfd_architecture @var{arch}, enum bfd_flavour @var{flavour}, enum gdb_osabi (*@var{sniffer})(bfd *@var{abfd}))
2655 Register the OS ABI file sniffer specified by @var{sniffer} for the
2656 BFD architecture/flavour pair specified by @var{arch} and @var{flavour}.
2657 If @var{arch} is @code{bfd_arch_unknown}, the sniffer is considered to
2658 be generic, and is allowed to examine @var{flavour}-flavoured files for
2662 @deftypefun enum gdb_osabi gdbarch_lookup_osabi (bfd *@var{abfd})
2663 Examine the file described by @var{abfd} to determine its OS ABI.
2664 The value @code{GDB_OSABI_UNKNOWN} is returned if the OS ABI cannot
2668 @deftypefun void gdbarch_init_osabi (struct gdbarch info @var{info}, struct gdbarch *@var{gdbarch}, enum gdb_osabi @var{osabi})
2669 Invoke the OS ABI handler corresponding to @var{osabi} to fine-tune the
2670 @code{gdbarch} structure specified by @var{gdbarch}. If a handler
2671 corresponding to @var{osabi} has not been registered for @var{gdbarch}'s
2672 architecture, a warning will be issued and the debugging session will continue
2673 with the defaults already established for @var{gdbarch}.
2676 @section Registers and Memory
2678 @value{GDBN}'s model of the target machine is rather simple.
2679 @value{GDBN} assumes the machine includes a bank of registers and a
2680 block of memory. Each register may have a different size.
2682 @value{GDBN} does not have a magical way to match up with the
2683 compiler's idea of which registers are which; however, it is critical
2684 that they do match up accurately. The only way to make this work is
2685 to get accurate information about the order that the compiler uses,
2686 and to reflect that in the @code{REGISTER_NAME} and related macros.
2688 @value{GDBN} can handle big-endian, little-endian, and bi-endian architectures.
2690 @section Pointers Are Not Always Addresses
2691 @cindex pointer representation
2692 @cindex address representation
2693 @cindex word-addressed machines
2694 @cindex separate data and code address spaces
2695 @cindex spaces, separate data and code address
2696 @cindex address spaces, separate data and code
2697 @cindex code pointers, word-addressed
2698 @cindex converting between pointers and addresses
2699 @cindex D10V addresses
2701 On almost all 32-bit architectures, the representation of a pointer is
2702 indistinguishable from the representation of some fixed-length number
2703 whose value is the byte address of the object pointed to. On such
2704 machines, the words ``pointer'' and ``address'' can be used interchangeably.
2705 However, architectures with smaller word sizes are often cramped for
2706 address space, so they may choose a pointer representation that breaks this
2707 identity, and allows a larger code address space.
2709 For example, the Renesas D10V is a 16-bit VLIW processor whose
2710 instructions are 32 bits long@footnote{Some D10V instructions are
2711 actually pairs of 16-bit sub-instructions. However, since you can't
2712 jump into the middle of such a pair, code addresses can only refer to
2713 full 32 bit instructions, which is what matters in this explanation.}.
2714 If the D10V used ordinary byte addresses to refer to code locations,
2715 then the processor would only be able to address 64kb of instructions.
2716 However, since instructions must be aligned on four-byte boundaries, the
2717 low two bits of any valid instruction's byte address are always
2718 zero---byte addresses waste two bits. So instead of byte addresses,
2719 the D10V uses word addresses---byte addresses shifted right two bits---to
2720 refer to code. Thus, the D10V can use 16-bit words to address 256kb of
2723 However, this means that code pointers and data pointers have different
2724 forms on the D10V. The 16-bit word @code{0xC020} refers to byte address
2725 @code{0xC020} when used as a data address, but refers to byte address
2726 @code{0x30080} when used as a code address.
2728 (The D10V also uses separate code and data address spaces, which also
2729 affects the correspondence between pointers and addresses, but we're
2730 going to ignore that here; this example is already too long.)
2732 To cope with architectures like this---the D10V is not the only
2733 one!---@value{GDBN} tries to distinguish between @dfn{addresses}, which are
2734 byte numbers, and @dfn{pointers}, which are the target's representation
2735 of an address of a particular type of data. In the example above,
2736 @code{0xC020} is the pointer, which refers to one of the addresses
2737 @code{0xC020} or @code{0x30080}, depending on the type imposed upon it.
2738 @value{GDBN} provides functions for turning a pointer into an address
2739 and vice versa, in the appropriate way for the current architecture.
2741 Unfortunately, since addresses and pointers are identical on almost all
2742 processors, this distinction tends to bit-rot pretty quickly. Thus,
2743 each time you port @value{GDBN} to an architecture which does
2744 distinguish between pointers and addresses, you'll probably need to
2745 clean up some architecture-independent code.
2747 Here are functions which convert between pointers and addresses:
2749 @deftypefun CORE_ADDR extract_typed_address (void *@var{buf}, struct type *@var{type})
2750 Treat the bytes at @var{buf} as a pointer or reference of type
2751 @var{type}, and return the address it represents, in a manner
2752 appropriate for the current architecture. This yields an address
2753 @value{GDBN} can use to read target memory, disassemble, etc. Note that
2754 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2757 For example, if the current architecture is the Intel x86, this function
2758 extracts a little-endian integer of the appropriate length from
2759 @var{buf} and returns it. However, if the current architecture is the
2760 D10V, this function will return a 16-bit integer extracted from
2761 @var{buf}, multiplied by four if @var{type} is a pointer to a function.
2763 If @var{type} is not a pointer or reference type, then this function
2764 will signal an internal error.
2767 @deftypefun CORE_ADDR store_typed_address (void *@var{buf}, struct type *@var{type}, CORE_ADDR @var{addr})
2768 Store the address @var{addr} in @var{buf}, in the proper format for a
2769 pointer of type @var{type} in the current architecture. Note that
2770 @var{buf} refers to a buffer in @value{GDBN}'s memory, not the
2773 For example, if the current architecture is the Intel x86, this function
2774 stores @var{addr} unmodified as a little-endian integer of the
2775 appropriate length in @var{buf}. However, if the current architecture
2776 is the D10V, this function divides @var{addr} by four if @var{type} is
2777 a pointer to a function, and then stores it in @var{buf}.
2779 If @var{type} is not a pointer or reference type, then this function
2780 will signal an internal error.
2783 @deftypefun CORE_ADDR value_as_address (struct value *@var{val})
2784 Assuming that @var{val} is a pointer, return the address it represents,
2785 as appropriate for the current architecture.
2787 This function actually works on integral values, as well as pointers.
2788 For pointers, it performs architecture-specific conversions as
2789 described above for @code{extract_typed_address}.
2792 @deftypefun CORE_ADDR value_from_pointer (struct type *@var{type}, CORE_ADDR @var{addr})
2793 Create and return a value representing a pointer of type @var{type} to
2794 the address @var{addr}, as appropriate for the current architecture.
2795 This function performs architecture-specific conversions as described
2796 above for @code{store_typed_address}.
2799 Here are some macros which architectures can define to indicate the
2800 relationship between pointers and addresses. These have default
2801 definitions, appropriate for architectures on which all pointers are
2802 simple unsigned byte addresses.
2804 @deftypefn {Target Macro} CORE_ADDR POINTER_TO_ADDRESS (struct type *@var{type}, char *@var{buf})
2805 Assume that @var{buf} holds a pointer of type @var{type}, in the
2806 appropriate format for the current architecture. Return the byte
2807 address the pointer refers to.
2809 This function may safely assume that @var{type} is either a pointer or a
2810 C@t{++} reference type.
2813 @deftypefn {Target Macro} void ADDRESS_TO_POINTER (struct type *@var{type}, char *@var{buf}, CORE_ADDR @var{addr})
2814 Store in @var{buf} a pointer of type @var{type} representing the address
2815 @var{addr}, in the appropriate format for the current architecture.
2817 This function may safely assume that @var{type} is either a pointer or a
2818 C@t{++} reference type.
2821 @section Address Classes
2822 @cindex address classes
2823 @cindex DW_AT_byte_size
2824 @cindex DW_AT_address_class
2826 Sometimes information about different kinds of addresses is available
2827 via the debug information. For example, some programming environments
2828 define addresses of several different sizes. If the debug information
2829 distinguishes these kinds of address classes through either the size
2830 info (e.g, @code{DW_AT_byte_size} in @w{DWARF 2}) or through an explicit
2831 address class attribute (e.g, @code{DW_AT_address_class} in @w{DWARF 2}), the
2832 following macros should be defined in order to disambiguate these
2833 types within @value{GDBN} as well as provide the added information to
2834 a @value{GDBN} user when printing type expressions.
2836 @deftypefn {Target Macro} int ADDRESS_CLASS_TYPE_FLAGS (int @var{byte_size}, int @var{dwarf2_addr_class})
2837 Returns the type flags needed to construct a pointer type whose size
2838 is @var{byte_size} and whose address class is @var{dwarf2_addr_class}.
2839 This function is normally called from within a symbol reader. See
2840 @file{dwarf2read.c}.
2843 @deftypefn {Target Macro} char *ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (int @var{type_flags})
2844 Given the type flags representing an address class qualifier, return
2847 @deftypefn {Target Macro} int ADDRESS_CLASS_NAME_to_TYPE_FLAGS (int @var{name}, int *var{type_flags_ptr})
2848 Given an address qualifier name, set the @code{int} refererenced by @var{type_flags_ptr} to the type flags
2849 for that address class qualifier.
2852 Since the need for address classes is rather rare, none of
2853 the address class macros defined by default. Predicate
2854 macros are provided to detect when they are defined.
2856 Consider a hypothetical architecture in which addresses are normally
2857 32-bits wide, but 16-bit addresses are also supported. Furthermore,
2858 suppose that the @w{DWARF 2} information for this architecture simply
2859 uses a @code{DW_AT_byte_size} value of 2 to indicate the use of one
2860 of these "short" pointers. The following functions could be defined
2861 to implement the address class macros:
2864 somearch_address_class_type_flags (int byte_size,
2865 int dwarf2_addr_class)
2868 return TYPE_FLAG_ADDRESS_CLASS_1;
2874 somearch_address_class_type_flags_to_name (int type_flags)
2876 if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
2883 somearch_address_class_name_to_type_flags (char *name,
2884 int *type_flags_ptr)
2886 if (strcmp (name, "short") == 0)
2888 *type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
2896 The qualifier @code{@@short} is used in @value{GDBN}'s type expressions
2897 to indicate the presence of one of these "short" pointers. E.g, if
2898 the debug information indicates that @code{short_ptr_var} is one of these
2899 short pointers, @value{GDBN} might show the following behavior:
2902 (gdb) ptype short_ptr_var
2903 type = int * @@short
2907 @section Raw and Virtual Register Representations
2908 @cindex raw register representation
2909 @cindex virtual register representation
2910 @cindex representations, raw and virtual registers
2912 @emph{Maintainer note: This section is pretty much obsolete. The
2913 functionality described here has largely been replaced by
2914 pseudo-registers and the mechanisms described in @ref{Target
2915 Architecture Definition, , Using Different Register and Memory Data
2916 Representations}. See also @uref{http://www.gnu.org/software/gdb/bugs/,
2917 Bug Tracking Database} and
2918 @uref{http://sources.redhat.com/gdb/current/ari/, ARI Index} for more
2919 up-to-date information.}
2921 Some architectures use one representation for a value when it lives in a
2922 register, but use a different representation when it lives in memory.
2923 In @value{GDBN}'s terminology, the @dfn{raw} representation is the one used in
2924 the target registers, and the @dfn{virtual} representation is the one
2925 used in memory, and within @value{GDBN} @code{struct value} objects.
2927 @emph{Maintainer note: Notice that the same mechanism is being used to
2928 both convert a register to a @code{struct value} and alternative
2931 For almost all data types on almost all architectures, the virtual and
2932 raw representations are identical, and no special handling is needed.
2933 However, they do occasionally differ. For example:
2937 The x86 architecture supports an 80-bit @code{long double} type. However, when
2938 we store those values in memory, they occupy twelve bytes: the
2939 floating-point number occupies the first ten, and the final two bytes
2940 are unused. This keeps the values aligned on four-byte boundaries,
2941 allowing more efficient access. Thus, the x86 80-bit floating-point
2942 type is the raw representation, and the twelve-byte loosely-packed
2943 arrangement is the virtual representation.
2946 Some 64-bit MIPS targets present 32-bit registers to @value{GDBN} as 64-bit
2947 registers, with garbage in their upper bits. @value{GDBN} ignores the top 32
2948 bits. Thus, the 64-bit form, with garbage in the upper 32 bits, is the
2949 raw representation, and the trimmed 32-bit representation is the
2950 virtual representation.
2953 In general, the raw representation is determined by the architecture, or
2954 @value{GDBN}'s interface to the architecture, while the virtual representation
2955 can be chosen for @value{GDBN}'s convenience. @value{GDBN}'s register file,
2956 @code{registers}, holds the register contents in raw format, and the
2957 @value{GDBN} remote protocol transmits register values in raw format.
2959 Your architecture may define the following macros to request
2960 conversions between the raw and virtual format:
2962 @deftypefn {Target Macro} int REGISTER_CONVERTIBLE (int @var{reg})
2963 Return non-zero if register number @var{reg}'s value needs different raw
2964 and virtual formats.
2966 You should not use @code{REGISTER_CONVERT_TO_VIRTUAL} for a register
2967 unless this macro returns a non-zero value for that register.
2970 @deftypefn {Target Macro} int DEPRECATED_REGISTER_RAW_SIZE (int @var{reg})
2971 The size of register number @var{reg}'s raw value. This is the number
2972 of bytes the register will occupy in @code{registers}, or in a @value{GDBN}
2973 remote protocol packet.
2976 @deftypefn {Target Macro} int DEPRECATED_REGISTER_VIRTUAL_SIZE (int @var{reg})
2977 The size of register number @var{reg}'s value, in its virtual format.
2978 This is the size a @code{struct value}'s buffer will have, holding that
2982 @deftypefn {Target Macro} struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int @var{reg})
2983 This is the type of the virtual representation of register number
2984 @var{reg}. Note that there is no need for a macro giving a type for the
2985 register's raw form; once the register's value has been obtained, @value{GDBN}
2986 always uses the virtual form.
2989 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_VIRTUAL (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
2990 Convert the value of register number @var{reg} to @var{type}, which
2991 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
2992 at @var{from} holds the register's value in raw format; the macro should
2993 convert the value to virtual format, and place it at @var{to}.
2995 Note that @code{REGISTER_CONVERT_TO_VIRTUAL} and
2996 @code{REGISTER_CONVERT_TO_RAW} take their @var{reg} and @var{type}
2997 arguments in different orders.
2999 You should only use @code{REGISTER_CONVERT_TO_VIRTUAL} with registers
3000 for which the @code{REGISTER_CONVERTIBLE} macro returns a non-zero
3004 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_RAW (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3005 Convert the value of register number @var{reg} to @var{type}, which
3006 should always be @code{DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})}. The buffer
3007 at @var{from} holds the register's value in raw format; the macro should
3008 convert the value to virtual format, and place it at @var{to}.
3010 Note that REGISTER_CONVERT_TO_VIRTUAL and REGISTER_CONVERT_TO_RAW take
3011 their @var{reg} and @var{type} arguments in different orders.
3015 @section Using Different Register and Memory Data Representations
3016 @cindex register representation
3017 @cindex memory representation
3018 @cindex representations, register and memory
3019 @cindex register data formats, converting
3020 @cindex @code{struct value}, converting register contents to
3022 @emph{Maintainer's note: The way GDB manipulates registers is undergoing
3023 significant change. Many of the macros and functions refered to in this
3024 section are likely to be subject to further revision. See
3025 @uref{http://sources.redhat.com/gdb/current/ari/, A.R. Index} and
3026 @uref{http://www.gnu.org/software/gdb/bugs, Bug Tracking Database} for
3027 further information. cagney/2002-05-06.}
3029 Some architectures can represent a data object in a register using a
3030 form that is different to the objects more normal memory representation.
3036 The Alpha architecture can represent 32 bit integer values in
3037 floating-point registers.
3040 The x86 architecture supports 80-bit floating-point registers. The
3041 @code{long double} data type occupies 96 bits in memory but only 80 bits
3042 when stored in a register.
3046 In general, the register representation of a data type is determined by
3047 the architecture, or @value{GDBN}'s interface to the architecture, while
3048 the memory representation is determined by the Application Binary
3051 For almost all data types on almost all architectures, the two
3052 representations are identical, and no special handling is needed.
3053 However, they do occasionally differ. Your architecture may define the
3054 following macros to request conversions between the register and memory
3055 representations of a data type:
3057 @deftypefn {Target Macro} int CONVERT_REGISTER_P (int @var{reg})
3058 Return non-zero if the representation of a data value stored in this
3059 register may be different to the representation of that same data value
3060 when stored in memory.
3062 When non-zero, the macros @code{REGISTER_TO_VALUE} and
3063 @code{VALUE_TO_REGISTER} are used to perform any necessary conversion.
3066 @deftypefn {Target Macro} void REGISTER_TO_VALUE (int @var{reg}, struct type *@var{type}, char *@var{from}, char *@var{to})
3067 Convert the value of register number @var{reg} to a data object of type
3068 @var{type}. The buffer at @var{from} holds the register's value in raw
3069 format; the converted value should be placed in the buffer at @var{to}.
3071 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
3072 their @var{reg} and @var{type} arguments in different orders.
3074 You should only use @code{REGISTER_TO_VALUE} with registers for which
3075 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
3078 @deftypefn {Target Macro} void VALUE_TO_REGISTER (struct type *@var{type}, int @var{reg}, char *@var{from}, char *@var{to})
3079 Convert a data value of type @var{type} to register number @var{reg}'
3082 Note that @code{REGISTER_TO_VALUE} and @code{VALUE_TO_REGISTER} take
3083 their @var{reg} and @var{type} arguments in different orders.
3085 You should only use @code{VALUE_TO_REGISTER} with registers for which
3086 the @code{CONVERT_REGISTER_P} macro returns a non-zero value.
3089 @deftypefn {Target Macro} void REGISTER_CONVERT_TO_TYPE (int @var{regnum}, struct type *@var{type}, char *@var{buf})
3090 See @file{mips-tdep.c}. It does not do what you want.
3094 @section Frame Interpretation
3096 @section Inferior Call Setup
3098 @section Compiler Characteristics
3100 @section Target Conditionals
3102 This section describes the macros that you can use to define the target
3107 @item ADDR_BITS_REMOVE (addr)
3108 @findex ADDR_BITS_REMOVE
3109 If a raw machine instruction address includes any bits that are not
3110 really part of the address, then define this macro to expand into an
3111 expression that zeroes those bits in @var{addr}. This is only used for
3112 addresses of instructions, and even then not in all contexts.
3114 For example, the two low-order bits of the PC on the Hewlett-Packard PA
3115 2.0 architecture contain the privilege level of the corresponding
3116 instruction. Since instructions must always be aligned on four-byte
3117 boundaries, the processor masks out these bits to generate the actual
3118 address of the instruction. ADDR_BITS_REMOVE should filter out these
3119 bits with an expression such as @code{((addr) & ~3)}.
3121 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS (@var{name}, @var{type_flags_ptr})
3122 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS
3123 If @var{name} is a valid address class qualifier name, set the @code{int}
3124 referenced by @var{type_flags_ptr} to the mask representing the qualifier
3125 and return 1. If @var{name} is not a valid address class qualifier name,
3128 The value for @var{type_flags_ptr} should be one of
3129 @code{TYPE_FLAG_ADDRESS_CLASS_1}, @code{TYPE_FLAG_ADDRESS_CLASS_2}, or
3130 possibly some combination of these values or'd together.
3131 @xref{Target Architecture Definition, , Address Classes}.
3133 @item ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P ()
3134 @findex ADDRESS_CLASS_NAME_TO_TYPE_FLAGS_P
3135 Predicate which indicates whether @code{ADDRESS_CLASS_NAME_TO_TYPE_FLAGS}
3138 @item ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
3139 @findex ADDRESS_CLASS_TYPE_FLAGS (@var{byte_size}, @var{dwarf2_addr_class})
3140 Given a pointers byte size (as described by the debug information) and
3141 the possible @code{DW_AT_address_class} value, return the type flags
3142 used by @value{GDBN} to represent this address class. The value
3143 returned should be one of @code{TYPE_FLAG_ADDRESS_CLASS_1},
3144 @code{TYPE_FLAG_ADDRESS_CLASS_2}, or possibly some combination of these
3145 values or'd together.
3146 @xref{Target Architecture Definition, , Address Classes}.
3148 @item ADDRESS_CLASS_TYPE_FLAGS_P ()
3149 @findex ADDRESS_CLASS_TYPE_FLAGS_P
3150 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS} has
3153 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME (@var{type_flags})
3154 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME
3155 Return the name of the address class qualifier associated with the type
3156 flags given by @var{type_flags}.
3158 @item ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P ()
3159 @findex ADDRESS_CLASS_TYPE_FLAGS_TO_NAME_P
3160 Predicate which indicates whether @code{ADDRESS_CLASS_TYPE_FLAGS_TO_NAME} has
3162 @xref{Target Architecture Definition, , Address Classes}.
3164 @item ADDRESS_TO_POINTER (@var{type}, @var{buf}, @var{addr})
3165 @findex ADDRESS_TO_POINTER
3166 Store in @var{buf} a pointer of type @var{type} representing the address
3167 @var{addr}, in the appropriate format for the current architecture.
3168 This macro may safely assume that @var{type} is either a pointer or a
3169 C@t{++} reference type.
3170 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3172 @item BELIEVE_PCC_PROMOTION
3173 @findex BELIEVE_PCC_PROMOTION
3174 Define if the compiler promotes a @code{short} or @code{char}
3175 parameter to an @code{int}, but still reports the parameter as its
3176 original type, rather than the promoted type.
3178 @item BITS_BIG_ENDIAN
3179 @findex BITS_BIG_ENDIAN
3180 Define this if the numbering of bits in the targets does @strong{not} match the
3181 endianness of the target byte order. A value of 1 means that the bits
3182 are numbered in a big-endian bit order, 0 means little-endian.
3186 This is the character array initializer for the bit pattern to put into
3187 memory where a breakpoint is set. Although it's common to use a trap
3188 instruction for a breakpoint, it's not required; for instance, the bit
3189 pattern could be an invalid instruction. The breakpoint must be no
3190 longer than the shortest instruction of the architecture.
3192 @code{BREAKPOINT} has been deprecated in favor of
3193 @code{BREAKPOINT_FROM_PC}.
3195 @item BIG_BREAKPOINT
3196 @itemx LITTLE_BREAKPOINT
3197 @findex LITTLE_BREAKPOINT
3198 @findex BIG_BREAKPOINT
3199 Similar to BREAKPOINT, but used for bi-endian targets.
3201 @code{BIG_BREAKPOINT} and @code{LITTLE_BREAKPOINT} have been deprecated in
3202 favor of @code{BREAKPOINT_FROM_PC}.
3204 @item DEPRECATED_REMOTE_BREAKPOINT
3205 @itemx DEPRECATED_LITTLE_REMOTE_BREAKPOINT
3206 @itemx DEPRECATED_BIG_REMOTE_BREAKPOINT
3207 @findex DEPRECATED_BIG_REMOTE_BREAKPOINT
3208 @findex DEPRECATED_LITTLE_REMOTE_BREAKPOINT
3209 @findex DEPRECATED_REMOTE_BREAKPOINT
3210 Specify the breakpoint instruction sequence for a remote target.
3211 @code{DEPRECATED_REMOTE_BREAKPOINT},
3212 @code{DEPRECATED_BIG_REMOTE_BREAKPOINT} and
3213 @code{DEPRECATED_LITTLE_REMOTE_BREAKPOINT} have been deprecated in
3214 favor of @code{BREAKPOINT_FROM_PC} (@pxref{BREAKPOINT_FROM_PC}).
3216 @item BREAKPOINT_FROM_PC (@var{pcptr}, @var{lenptr})
3217 @findex BREAKPOINT_FROM_PC
3218 @anchor{BREAKPOINT_FROM_PC} Use the program counter to determine the
3219 contents and size of a breakpoint instruction. It returns a pointer to
3220 a string of bytes that encode a breakpoint instruction, stores the
3221 length of the string to @code{*@var{lenptr}}, and adjusts the program
3222 counter (if necessary) to point to the actual memory location where the
3223 breakpoint should be inserted.
3225 Although it is common to use a trap instruction for a breakpoint, it's
3226 not required; for instance, the bit pattern could be an invalid
3227 instruction. The breakpoint must be no longer than the shortest
3228 instruction of the architecture.
3230 Replaces all the other @var{BREAKPOINT} macros.
3232 @item MEMORY_INSERT_BREAKPOINT (@var{addr}, @var{contents_cache})
3233 @itemx MEMORY_REMOVE_BREAKPOINT (@var{addr}, @var{contents_cache})
3234 @findex MEMORY_REMOVE_BREAKPOINT
3235 @findex MEMORY_INSERT_BREAKPOINT
3236 Insert or remove memory based breakpoints. Reasonable defaults
3237 (@code{default_memory_insert_breakpoint} and
3238 @code{default_memory_remove_breakpoint} respectively) have been
3239 provided so that it is not necessary to define these for most
3240 architectures. Architectures which may want to define
3241 @code{MEMORY_INSERT_BREAKPOINT} and @code{MEMORY_REMOVE_BREAKPOINT} will
3242 likely have instructions that are oddly sized or are not stored in a
3243 conventional manner.
3245 It may also be desirable (from an efficiency standpoint) to define
3246 custom breakpoint insertion and removal routines if
3247 @code{BREAKPOINT_FROM_PC} needs to read the target's memory for some
3250 @item ADJUST_BREAKPOINT_ADDRESS (@var{address})
3251 @findex ADJUST_BREAKPOINT_ADDRESS
3252 @cindex breakpoint address adjusted
3253 Given an address at which a breakpoint is desired, return a breakpoint
3254 address adjusted to account for architectural constraints on
3255 breakpoint placement. This method is not needed by most targets.
3257 The FR-V target (see @file{frv-tdep.c}) requires this method.
3258 The FR-V is a VLIW architecture in which a number of RISC-like
3259 instructions are grouped (packed) together into an aggregate
3260 instruction or instruction bundle. When the processor executes
3261 one of these bundles, the component instructions are executed
3264 In the course of optimization, the compiler may group instructions
3265 from distinct source statements into the same bundle. The line number
3266 information associated with one of the latter statements will likely
3267 refer to some instruction other than the first one in the bundle. So,
3268 if the user attempts to place a breakpoint on one of these latter
3269 statements, @value{GDBN} must be careful to @emph{not} place the break
3270 instruction on any instruction other than the first one in the bundle.
3271 (Remember though that the instructions within a bundle execute
3272 in parallel, so the @emph{first} instruction is the instruction
3273 at the lowest address and has nothing to do with execution order.)
3275 The FR-V's @code{ADJUST_BREAKPOINT_ADDRESS} method will adjust a
3276 breakpoint's address by scanning backwards for the beginning of
3277 the bundle, returning the address of the bundle.
3279 Since the adjustment of a breakpoint may significantly alter a user's
3280 expectation, @value{GDBN} prints a warning when an adjusted breakpoint
3281 is initially set and each time that that breakpoint is hit.
3283 @item CALL_DUMMY_LOCATION
3284 @findex CALL_DUMMY_LOCATION
3285 See the file @file{inferior.h}.
3287 This method has been replaced by @code{push_dummy_code}
3288 (@pxref{push_dummy_code}).
3290 @item CANNOT_FETCH_REGISTER (@var{regno})
3291 @findex CANNOT_FETCH_REGISTER
3292 A C expression that should be nonzero if @var{regno} cannot be fetched
3293 from an inferior process. This is only relevant if
3294 @code{FETCH_INFERIOR_REGISTERS} is not defined.
3296 @item CANNOT_STORE_REGISTER (@var{regno})
3297 @findex CANNOT_STORE_REGISTER
3298 A C expression that should be nonzero if @var{regno} should not be
3299 written to the target. This is often the case for program counters,
3300 status words, and other special registers. If this is not defined,
3301 @value{GDBN} will assume that all registers may be written.
3303 @item int CONVERT_REGISTER_P(@var{regnum})
3304 @findex CONVERT_REGISTER_P
3305 Return non-zero if register @var{regnum} can represent data values in a
3307 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3309 @item DECR_PC_AFTER_BREAK
3310 @findex DECR_PC_AFTER_BREAK
3311 Define this to be the amount by which to decrement the PC after the
3312 program encounters a breakpoint. This is often the number of bytes in
3313 @code{BREAKPOINT}, though not always. For most targets this value will be 0.
3315 @item DISABLE_UNSETTABLE_BREAK (@var{addr})
3316 @findex DISABLE_UNSETTABLE_BREAK
3317 If defined, this should evaluate to 1 if @var{addr} is in a shared
3318 library in which breakpoints cannot be set and so should be disabled.
3320 @item PRINT_FLOAT_INFO()
3321 @findex PRINT_FLOAT_INFO
3322 If defined, then the @samp{info float} command will print information about
3323 the processor's floating point unit.
3325 @item print_registers_info (@var{gdbarch}, @var{frame}, @var{regnum}, @var{all})
3326 @findex print_registers_info
3327 If defined, pretty print the value of the register @var{regnum} for the
3328 specified @var{frame}. If the value of @var{regnum} is -1, pretty print
3329 either all registers (@var{all} is non zero) or a select subset of
3330 registers (@var{all} is zero).
3332 The default method prints one register per line, and if @var{all} is
3333 zero omits floating-point registers.
3335 @item PRINT_VECTOR_INFO()
3336 @findex PRINT_VECTOR_INFO
3337 If defined, then the @samp{info vector} command will call this function
3338 to print information about the processor's vector unit.
3340 By default, the @samp{info vector} command will print all vector
3341 registers (the register's type having the vector attribute).
3343 @item DWARF_REG_TO_REGNUM
3344 @findex DWARF_REG_TO_REGNUM
3345 Convert DWARF register number into @value{GDBN} regnum. If not defined,
3346 no conversion will be performed.
3348 @item DWARF2_REG_TO_REGNUM
3349 @findex DWARF2_REG_TO_REGNUM
3350 Convert DWARF2 register number into @value{GDBN} regnum. If not
3351 defined, no conversion will be performed.
3353 @item ECOFF_REG_TO_REGNUM
3354 @findex ECOFF_REG_TO_REGNUM
3355 Convert ECOFF register number into @value{GDBN} regnum. If not defined,
3356 no conversion will be performed.
3358 @item END_OF_TEXT_DEFAULT
3359 @findex END_OF_TEXT_DEFAULT
3360 This is an expression that should designate the end of the text section.
3363 @item EXTRACT_RETURN_VALUE(@var{type}, @var{regbuf}, @var{valbuf})
3364 @findex EXTRACT_RETURN_VALUE
3365 Define this to extract a function's return value of type @var{type} from
3366 the raw register state @var{regbuf} and copy that, in virtual format,
3369 This method has been deprecated in favour of @code{gdbarch_return_value}
3370 (@pxref{gdbarch_return_value}).
3372 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(@var{regbuf})
3373 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS
3374 @anchor{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}
3375 When defined, extract from the array @var{regbuf} (containing the raw
3376 register state) the @code{CORE_ADDR} at which a function should return
3377 its structure value.
3379 @xref{gdbarch_return_value}.
3381 @item DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P()
3382 @findex DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P
3383 Predicate for @code{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}.
3385 @item DEPRECATED_FP_REGNUM
3386 @findex DEPRECATED_FP_REGNUM
3387 If the virtual frame pointer is kept in a register, then define this
3388 macro to be the number (greater than or equal to zero) of that register.
3390 This should only need to be defined if @code{DEPRECATED_TARGET_READ_FP}
3393 @item DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(@var{fi})
3394 @findex DEPRECATED_FRAMELESS_FUNCTION_INVOCATION
3395 Define this to an expression that returns 1 if the function invocation
3396 represented by @var{fi} does not have a stack frame associated with it.
3399 @item frame_align (@var{address})
3400 @anchor{frame_align}
3402 Define this to adjust @var{address} so that it meets the alignment
3403 requirements for the start of a new stack frame. A stack frame's
3404 alignment requirements are typically stronger than a target processors
3405 stack alignment requirements (@pxref{DEPRECATED_STACK_ALIGN}).
3407 This function is used to ensure that, when creating a dummy frame, both
3408 the initial stack pointer and (if needed) the address of the return
3409 value are correctly aligned.
3411 Unlike @code{DEPRECATED_STACK_ALIGN}, this function always adjusts the
3412 address in the direction of stack growth.
3414 By default, no frame based stack alignment is performed.
3416 @item int frame_red_zone_size
3418 The number of bytes, beyond the innermost-stack-address, reserved by the
3419 @sc{abi}. A function is permitted to use this scratch area (instead of
3420 allocating extra stack space).
3422 When performing an inferior function call, to ensure that it does not
3423 modify this area, @value{GDBN} adjusts the innermost-stack-address by
3424 @var{frame_red_zone_size} bytes before pushing parameters onto the
3427 By default, zero bytes are allocated. The value must be aligned
3428 (@pxref{frame_align}).
3430 The @sc{amd64} (nee x86-64) @sc{abi} documentation refers to the
3431 @emph{red zone} when describing this scratch area.
3434 @item DEPRECATED_FRAME_CHAIN(@var{frame})
3435 @findex DEPRECATED_FRAME_CHAIN
3436 Given @var{frame}, return a pointer to the calling frame.
3438 @item DEPRECATED_FRAME_CHAIN_VALID(@var{chain}, @var{thisframe})
3439 @findex DEPRECATED_FRAME_CHAIN_VALID
3440 Define this to be an expression that returns zero if the given frame is an
3441 outermost frame, with no caller, and nonzero otherwise. Most normal
3442 situations can be handled without defining this macro, including @code{NULL}
3443 chain pointers, dummy frames, and frames whose PC values are inside the
3444 startup file (e.g.@: @file{crt0.o}), inside @code{main}, or inside
3447 @item DEPRECATED_FRAME_INIT_SAVED_REGS(@var{frame})
3448 @findex DEPRECATED_FRAME_INIT_SAVED_REGS
3449 See @file{frame.h}. Determines the address of all registers in the
3450 current stack frame storing each in @code{frame->saved_regs}. Space for
3451 @code{frame->saved_regs} shall be allocated by
3452 @code{DEPRECATED_FRAME_INIT_SAVED_REGS} using
3453 @code{frame_saved_regs_zalloc}.
3455 @code{FRAME_FIND_SAVED_REGS} is deprecated.
3457 @item FRAME_NUM_ARGS (@var{fi})
3458 @findex FRAME_NUM_ARGS
3459 For the frame described by @var{fi} return the number of arguments that
3460 are being passed. If the number of arguments is not known, return
3463 @item DEPRECATED_FRAME_SAVED_PC(@var{frame})
3464 @findex DEPRECATED_FRAME_SAVED_PC
3465 @anchor{DEPRECATED_FRAME_SAVED_PC} Given @var{frame}, return the pc
3466 saved there. This is the return address.
3468 This method is deprecated. @xref{unwind_pc}.
3470 @item CORE_ADDR unwind_pc (struct frame_info *@var{this_frame})
3472 @anchor{unwind_pc} Return the instruction address, in @var{this_frame}'s
3473 caller, at which execution will resume after @var{this_frame} returns.
3474 This is commonly refered to as the return address.
3476 The implementation, which must be frame agnostic (work with any frame),
3477 is typically no more than:
3481 frame_unwind_unsigned_register (this_frame, D10V_PC_REGNUM, &pc);
3482 return d10v_make_iaddr (pc);
3486 @xref{DEPRECATED_FRAME_SAVED_PC}, which this method replaces.
3488 @item CORE_ADDR unwind_sp (struct frame_info *@var{this_frame})
3490 @anchor{unwind_sp} Return the frame's inner most stack address. This is
3491 commonly refered to as the frame's @dfn{stack pointer}.
3493 The implementation, which must be frame agnostic (work with any frame),
3494 is typically no more than:
3498 frame_unwind_unsigned_register (this_frame, D10V_SP_REGNUM, &sp);
3499 return d10v_make_daddr (sp);
3503 @xref{TARGET_READ_SP}, which this method replaces.
3505 @item FUNCTION_EPILOGUE_SIZE
3506 @findex FUNCTION_EPILOGUE_SIZE
3507 For some COFF targets, the @code{x_sym.x_misc.x_fsize} field of the
3508 function end symbol is 0. For such targets, you must define
3509 @code{FUNCTION_EPILOGUE_SIZE} to expand into the standard size of a
3510 function's epilogue.
3512 @item DEPRECATED_FUNCTION_START_OFFSET
3513 @findex DEPRECATED_FUNCTION_START_OFFSET
3514 An integer, giving the offset in bytes from a function's address (as
3515 used in the values of symbols, function pointers, etc.), and the
3516 function's first genuine instruction.
3518 This is zero on almost all machines: the function's address is usually
3519 the address of its first instruction. However, on the VAX, for
3520 example, each function starts with two bytes containing a bitmask
3521 indicating which registers to save upon entry to the function. The
3522 VAX @code{call} instructions check this value, and save the
3523 appropriate registers automatically. Thus, since the offset from the
3524 function's address to its first instruction is two bytes,
3525 @code{DEPRECATED_FUNCTION_START_OFFSET} would be 2 on the VAX.
3527 @item GCC_COMPILED_FLAG_SYMBOL
3528 @itemx GCC2_COMPILED_FLAG_SYMBOL
3529 @findex GCC2_COMPILED_FLAG_SYMBOL
3530 @findex GCC_COMPILED_FLAG_SYMBOL
3531 If defined, these are the names of the symbols that @value{GDBN} will
3532 look for to detect that GCC compiled the file. The default symbols
3533 are @code{gcc_compiled.} and @code{gcc2_compiled.},
3534 respectively. (Currently only defined for the Delta 68.)
3536 @item @value{GDBN}_MULTI_ARCH
3537 @findex @value{GDBN}_MULTI_ARCH
3538 If defined and non-zero, enables support for multiple architectures
3539 within @value{GDBN}.
3541 This support can be enabled at two levels. At level one, only
3542 definitions for previously undefined macros are provided; at level two,
3543 a multi-arch definition of all architecture dependent macros will be
3546 @item @value{GDBN}_TARGET_IS_HPPA
3547 @findex @value{GDBN}_TARGET_IS_HPPA
3548 This determines whether horrible kludge code in @file{dbxread.c} and
3549 @file{partial-stab.h} is used to mangle multiple-symbol-table files from
3550 HPPA's. This should all be ripped out, and a scheme like @file{elfread.c}
3553 @item GET_LONGJMP_TARGET
3554 @findex GET_LONGJMP_TARGET
3555 For most machines, this is a target-dependent parameter. On the
3556 DECstation and the Iris, this is a native-dependent parameter, since
3557 the header file @file{setjmp.h} is needed to define it.
3559 This macro determines the target PC address that @code{longjmp} will jump to,
3560 assuming that we have just stopped at a @code{longjmp} breakpoint. It takes a
3561 @code{CORE_ADDR *} as argument, and stores the target PC value through this
3562 pointer. It examines the current state of the machine as needed.
3564 @item DEPRECATED_GET_SAVED_REGISTER
3565 @findex DEPRECATED_GET_SAVED_REGISTER
3566 Define this if you need to supply your own definition for the function
3567 @code{DEPRECATED_GET_SAVED_REGISTER}.
3569 @item DEPRECATED_IBM6000_TARGET
3570 @findex DEPRECATED_IBM6000_TARGET
3571 Shows that we are configured for an IBM RS/6000 system. This
3572 conditional should be eliminated (FIXME) and replaced by
3573 feature-specific macros. It was introduced in a haste and we are
3574 repenting at leisure.
3576 @item I386_USE_GENERIC_WATCHPOINTS
3577 An x86-based target can define this to use the generic x86 watchpoint
3578 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
3580 @item SYMBOLS_CAN_START_WITH_DOLLAR
3581 @findex SYMBOLS_CAN_START_WITH_DOLLAR
3582 Some systems have routines whose names start with @samp{$}. Giving this
3583 macro a non-zero value tells @value{GDBN}'s expression parser to check for such
3584 routines when parsing tokens that begin with @samp{$}.
3586 On HP-UX, certain system routines (millicode) have names beginning with
3587 @samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
3588 routine that handles inter-space procedure calls on PA-RISC.
3590 @item DEPRECATED_INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
3591 @findex DEPRECATED_INIT_EXTRA_FRAME_INFO
3592 If additional information about the frame is required this should be
3593 stored in @code{frame->extra_info}. Space for @code{frame->extra_info}
3594 is allocated using @code{frame_extra_info_zalloc}.
3596 @item DEPRECATED_INIT_FRAME_PC (@var{fromleaf}, @var{prev})
3597 @findex DEPRECATED_INIT_FRAME_PC
3598 This is a C statement that sets the pc of the frame pointed to by
3599 @var{prev}. [By default...]
3601 @item INNER_THAN (@var{lhs}, @var{rhs})
3603 Returns non-zero if stack address @var{lhs} is inner than (nearer to the
3604 stack top) stack address @var{rhs}. Define this as @code{lhs < rhs} if
3605 the target's stack grows downward in memory, or @code{lhs > rsh} if the
3608 @item gdbarch_in_function_epilogue_p (@var{gdbarch}, @var{pc})
3609 @findex gdbarch_in_function_epilogue_p
3610 Returns non-zero if the given @var{pc} is in the epilogue of a function.
3611 The epilogue of a function is defined as the part of a function where
3612 the stack frame of the function already has been destroyed up to the
3613 final `return from function call' instruction.
3615 @item DEPRECATED_SIGTRAMP_START (@var{pc})
3616 @findex DEPRECATED_SIGTRAMP_START
3617 @itemx DEPRECATED_SIGTRAMP_END (@var{pc})
3618 @findex DEPRECATED_SIGTRAMP_END
3619 Define these to be the start and end address of the @code{sigtramp} for the
3620 given @var{pc}. On machines where the address is just a compile time
3621 constant, the macro expansion will typically just ignore the supplied
3624 @item IN_SOLIB_CALL_TRAMPOLINE (@var{pc}, @var{name})
3625 @findex IN_SOLIB_CALL_TRAMPOLINE
3626 Define this to evaluate to nonzero if the program is stopped in the
3627 trampoline that connects to a shared library.
3629 @item IN_SOLIB_RETURN_TRAMPOLINE (@var{pc}, @var{name})
3630 @findex IN_SOLIB_RETURN_TRAMPOLINE
3631 Define this to evaluate to nonzero if the program is stopped in the
3632 trampoline that returns from a shared library.
3634 @item IN_SOLIB_DYNSYM_RESOLVE_CODE (@var{pc})
3635 @findex IN_SOLIB_DYNSYM_RESOLVE_CODE
3636 Define this to evaluate to nonzero if the program is stopped in the
3639 @item SKIP_SOLIB_RESOLVER (@var{pc})
3640 @findex SKIP_SOLIB_RESOLVER
3641 Define this to evaluate to the (nonzero) address at which execution
3642 should continue to get past the dynamic linker's symbol resolution
3643 function. A zero value indicates that it is not important or necessary
3644 to set a breakpoint to get through the dynamic linker and that single
3645 stepping will suffice.
3647 @item INTEGER_TO_ADDRESS (@var{type}, @var{buf})
3648 @findex INTEGER_TO_ADDRESS
3649 @cindex converting integers to addresses
3650 Define this when the architecture needs to handle non-pointer to address
3651 conversions specially. Converts that value to an address according to
3652 the current architectures conventions.
3654 @emph{Pragmatics: When the user copies a well defined expression from
3655 their source code and passes it, as a parameter, to @value{GDBN}'s
3656 @code{print} command, they should get the same value as would have been
3657 computed by the target program. Any deviation from this rule can cause
3658 major confusion and annoyance, and needs to be justified carefully. In
3659 other words, @value{GDBN} doesn't really have the freedom to do these
3660 conversions in clever and useful ways. It has, however, been pointed
3661 out that users aren't complaining about how @value{GDBN} casts integers
3662 to pointers; they are complaining that they can't take an address from a
3663 disassembly listing and give it to @code{x/i}. Adding an architecture
3664 method like @code{INTEGER_TO_ADDRESS} certainly makes it possible for
3665 @value{GDBN} to ``get it right'' in all circumstances.}
3667 @xref{Target Architecture Definition, , Pointers Are Not Always
3670 @item NO_HIF_SUPPORT
3671 @findex NO_HIF_SUPPORT
3672 (Specific to the a29k.)
3674 @item POINTER_TO_ADDRESS (@var{type}, @var{buf})
3675 @findex POINTER_TO_ADDRESS
3676 Assume that @var{buf} holds a pointer of type @var{type}, in the
3677 appropriate format for the current architecture. Return the byte
3678 address the pointer refers to.
3679 @xref{Target Architecture Definition, , Pointers Are Not Always Addresses}.
3681 @item REGISTER_CONVERTIBLE (@var{reg})
3682 @findex REGISTER_CONVERTIBLE
3683 Return non-zero if @var{reg} uses different raw and virtual formats.
3684 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3686 @item REGISTER_TO_VALUE(@var{regnum}, @var{type}, @var{from}, @var{to})
3687 @findex REGISTER_TO_VALUE
3688 Convert the raw contents of register @var{regnum} into a value of type
3690 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
3692 @item DEPRECATED_REGISTER_RAW_SIZE (@var{reg})
3693 @findex DEPRECATED_REGISTER_RAW_SIZE
3694 Return the raw size of @var{reg}; defaults to the size of the register's
3696 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3698 @item register_reggroup_p (@var{gdbarch}, @var{regnum}, @var{reggroup})
3699 @findex register_reggroup_p
3700 @cindex register groups
3701 Return non-zero if register @var{regnum} is a member of the register
3702 group @var{reggroup}.
3704 By default, registers are grouped as follows:
3707 @item float_reggroup
3708 Any register with a valid name and a floating-point type.
3709 @item vector_reggroup
3710 Any register with a valid name and a vector type.
3711 @item general_reggroup
3712 Any register with a valid name and a type other than vector or
3713 floating-point. @samp{float_reggroup}.
3715 @itemx restore_reggroup
3717 Any register with a valid name.
3720 @item DEPRECATED_REGISTER_VIRTUAL_SIZE (@var{reg})
3721 @findex DEPRECATED_REGISTER_VIRTUAL_SIZE
3722 Return the virtual size of @var{reg}; defaults to the size of the
3723 register's virtual type.
3724 Return the virtual size of @var{reg}.
3725 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3727 @item DEPRECATED_REGISTER_VIRTUAL_TYPE (@var{reg})
3728 @findex REGISTER_VIRTUAL_TYPE
3729 Return the virtual type of @var{reg}.
3730 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3732 @item struct type *register_type (@var{gdbarch}, @var{reg})
3733 @findex register_type
3734 If defined, return the type of register @var{reg}. This function
3735 superseeds @code{DEPRECATED_REGISTER_VIRTUAL_TYPE}. @xref{Target Architecture
3736 Definition, , Raw and Virtual Register Representations}.
3738 @item REGISTER_CONVERT_TO_VIRTUAL(@var{reg}, @var{type}, @var{from}, @var{to})
3739 @findex REGISTER_CONVERT_TO_VIRTUAL
3740 Convert the value of register @var{reg} from its raw form to its virtual
3742 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3744 @item REGISTER_CONVERT_TO_RAW(@var{type}, @var{reg}, @var{from}, @var{to})
3745 @findex REGISTER_CONVERT_TO_RAW
3746 Convert the value of register @var{reg} from its virtual form to its raw
3748 @xref{Target Architecture Definition, , Raw and Virtual Register Representations}.
3750 @item const struct regset *regset_from_core_section (struct gdbarch * @var{gdbarch}, const char * @var{sect_name}, size_t @var{sect_size})
3751 @findex regset_from_core_section
3752 Return the appropriate register set for a core file section with name
3753 @var{sect_name} and size @var{sect_size}.
3755 @item SOFTWARE_SINGLE_STEP_P()
3756 @findex SOFTWARE_SINGLE_STEP_P
3757 Define this as 1 if the target does not have a hardware single-step
3758 mechanism. The macro @code{SOFTWARE_SINGLE_STEP} must also be defined.
3760 @item SOFTWARE_SINGLE_STEP(@var{signal}, @var{insert_breapoints_p})
3761 @findex SOFTWARE_SINGLE_STEP
3762 A function that inserts or removes (depending on
3763 @var{insert_breapoints_p}) breakpoints at each possible destinations of
3764 the next instruction. See @file{sparc-tdep.c} and @file{rs6000-tdep.c}
3767 @item SOFUN_ADDRESS_MAYBE_MISSING
3768 @findex SOFUN_ADDRESS_MAYBE_MISSING
3769 Somebody clever observed that, the more actual addresses you have in the
3770 debug information, the more time the linker has to spend relocating
3771 them. So whenever there's some other way the debugger could find the
3772 address it needs, you should omit it from the debug info, to make
3775 @code{SOFUN_ADDRESS_MAYBE_MISSING} indicates that a particular set of
3776 hacks of this sort are in use, affecting @code{N_SO} and @code{N_FUN}
3777 entries in stabs-format debugging information. @code{N_SO} stabs mark
3778 the beginning and ending addresses of compilation units in the text
3779 segment. @code{N_FUN} stabs mark the starts and ends of functions.
3781 @code{SOFUN_ADDRESS_MAYBE_MISSING} means two things:
3785 @code{N_FUN} stabs have an address of zero. Instead, you should find the
3786 addresses where the function starts by taking the function name from
3787 the stab, and then looking that up in the minsyms (the
3788 linker/assembler symbol table). In other words, the stab has the
3789 name, and the linker/assembler symbol table is the only place that carries
3793 @code{N_SO} stabs have an address of zero, too. You just look at the
3794 @code{N_FUN} stabs that appear before and after the @code{N_SO} stab,
3795 and guess the starting and ending addresses of the compilation unit from
3799 @item PC_LOAD_SEGMENT
3800 @findex PC_LOAD_SEGMENT
3801 If defined, print information about the load segment for the program
3802 counter. (Defined only for the RS/6000.)
3806 If the program counter is kept in a register, then define this macro to
3807 be the number (greater than or equal to zero) of that register.
3809 This should only need to be defined if @code{TARGET_READ_PC} and
3810 @code{TARGET_WRITE_PC} are not defined.
3813 @findex PARM_BOUNDARY
3814 If non-zero, round arguments to a boundary of this many bits before
3815 pushing them on the stack.
3817 @item stabs_argument_has_addr (@var{gdbarch}, @var{type})
3818 @findex stabs_argument_has_addr
3819 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3820 @anchor{stabs_argument_has_addr} Define this to return nonzero if a
3821 function argument of type @var{type} is passed by reference instead of
3824 This method replaces @code{DEPRECATED_REG_STRUCT_HAS_ADDR}
3825 (@pxref{DEPRECATED_REG_STRUCT_HAS_ADDR}).
3827 @item PROCESS_LINENUMBER_HOOK
3828 @findex PROCESS_LINENUMBER_HOOK
3829 A hook defined for XCOFF reading.
3831 @item PROLOGUE_FIRSTLINE_OVERLAP
3832 @findex PROLOGUE_FIRSTLINE_OVERLAP
3833 (Only used in unsupported Convex configuration.)
3837 If defined, this is the number of the processor status register. (This
3838 definition is only used in generic code when parsing "$ps".)
3840 @item DEPRECATED_POP_FRAME
3841 @findex DEPRECATED_POP_FRAME
3843 If defined, used by @code{frame_pop} to remove a stack frame. This
3844 method has been superseeded by generic code.
3846 @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})
3847 @findex push_dummy_call
3848 @findex DEPRECATED_PUSH_ARGUMENTS.
3849 @anchor{push_dummy_call} Define this to push the dummy frame's call to
3850 the inferior function onto the stack. In addition to pushing
3851 @var{nargs}, the code should push @var{struct_addr} (when
3852 @var{struct_return}), and the return address (@var{bp_addr}).
3854 @var{function} is a pointer to a @code{struct value}; on architectures that use
3855 function descriptors, this contains the function descriptor value.
3857 Returns the updated top-of-stack pointer.
3859 This method replaces @code{DEPRECATED_PUSH_ARGUMENTS}.
3861 @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})
3862 @findex push_dummy_code
3863 @anchor{push_dummy_code} Given a stack based call dummy, push the
3864 instruction sequence (including space for a breakpoint) to which the
3865 called function should return.
3867 Set @var{bp_addr} to the address at which the breakpoint instruction
3868 should be inserted, @var{real_pc} to the resume address when starting
3869 the call sequence, and return the updated inner-most stack address.
3871 By default, the stack is grown sufficient to hold a frame-aligned
3872 (@pxref{frame_align}) breakpoint, @var{bp_addr} is set to the address
3873 reserved for that breakpoint, and @var{real_pc} set to @var{funaddr}.
3875 This method replaces @code{CALL_DUMMY_LOCATION},
3876 @code{DEPRECATED_REGISTER_SIZE}.
3878 @item REGISTER_NAME(@var{i})
3879 @findex REGISTER_NAME
3880 Return the name of register @var{i} as a string. May return @code{NULL}
3881 or @code{NUL} to indicate that register @var{i} is not valid.
3883 @item DEPRECATED_REG_STRUCT_HAS_ADDR (@var{gcc_p}, @var{type})
3884 @findex DEPRECATED_REG_STRUCT_HAS_ADDR
3885 @anchor{DEPRECATED_REG_STRUCT_HAS_ADDR}Define this to return 1 if the
3886 given type will be passed by pointer rather than directly.
3888 This method has been replaced by @code{stabs_argument_has_addr}
3889 (@pxref{stabs_argument_has_addr}).
3891 @item SAVE_DUMMY_FRAME_TOS (@var{sp})
3892 @findex SAVE_DUMMY_FRAME_TOS
3893 @anchor{SAVE_DUMMY_FRAME_TOS} Used in @samp{call_function_by_hand} to
3894 notify the target dependent code of the top-of-stack value that will be
3895 passed to the the inferior code. This is the value of the @code{SP}
3896 after both the dummy frame and space for parameters/results have been
3897 allocated on the stack. @xref{unwind_dummy_id}.
3899 @item SDB_REG_TO_REGNUM
3900 @findex SDB_REG_TO_REGNUM
3901 Define this to convert sdb register numbers into @value{GDBN} regnums. If not
3902 defined, no conversion will be done.
3904 @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})
3905 @findex gdbarch_return_value
3906 @anchor{gdbarch_return_value} Given a function with a return-value of
3907 type @var{rettype}, return which return-value convention that function
3910 @value{GDBN} currently recognizes two function return-value conventions:
3911 @code{RETURN_VALUE_REGISTER_CONVENTION} where the return value is found
3912 in registers; and @code{RETURN_VALUE_STRUCT_CONVENTION} where the return
3913 value is found in memory and the address of that memory location is
3914 passed in as the function's first parameter.
3916 If the register convention is being used, and @var{writebuf} is
3917 non-@code{NULL}, also copy the return-value in @var{writebuf} into
3920 If the register convention is being used, and @var{readbuf} is
3921 non-@code{NULL}, also copy the return value from @var{regcache} into
3922 @var{readbuf} (@var{regcache} contains a copy of the registers from the
3923 just returned function).
3925 @xref{DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS}, for a description of how
3926 return-values that use the struct convention are handled.
3928 @emph{Maintainer note: This method replaces separate predicate, extract,
3929 store methods. By having only one method, the logic needed to determine
3930 the return-value convention need only be implemented in one place. If
3931 @value{GDBN} were written in an @sc{oo} language, this method would
3932 instead return an object that knew how to perform the register
3933 return-value extract and store.}
3935 @emph{Maintainer note: This method does not take a @var{gcc_p}
3936 parameter, and such a parameter should not be added. If an architecture
3937 that requires per-compiler or per-function information be identified,
3938 then the replacement of @var{rettype} with @code{struct value}
3939 @var{function} should be persued.}
3941 @emph{Maintainer note: The @var{regcache} parameter limits this methods
3942 to the inner most frame. While replacing @var{regcache} with a
3943 @code{struct frame_info} @var{frame} parameter would remove that
3944 limitation there has yet to be a demonstrated need for such a change.}
3946 @item SKIP_PERMANENT_BREAKPOINT
3947 @findex SKIP_PERMANENT_BREAKPOINT
3948 Advance the inferior's PC past a permanent breakpoint. @value{GDBN} normally
3949 steps over a breakpoint by removing it, stepping one instruction, and
3950 re-inserting the breakpoint. However, permanent breakpoints are
3951 hardwired into the inferior, and can't be removed, so this strategy
3952 doesn't work. Calling @code{SKIP_PERMANENT_BREAKPOINT} adjusts the processor's
3953 state so that execution will resume just after the breakpoint. This
3954 macro does the right thing even when the breakpoint is in the delay slot
3955 of a branch or jump.
3957 @item SKIP_PROLOGUE (@var{pc})
3958 @findex SKIP_PROLOGUE
3959 A C expression that returns the address of the ``real'' code beyond the
3960 function entry prologue found at @var{pc}.
3962 @item SKIP_TRAMPOLINE_CODE (@var{pc})
3963 @findex SKIP_TRAMPOLINE_CODE
3964 If the target machine has trampoline code that sits between callers and
3965 the functions being called, then define this macro to return a new PC
3966 that is at the start of the real function.
3970 If the stack-pointer is kept in a register, then define this macro to be
3971 the number (greater than or equal to zero) of that register, or -1 if
3972 there is no such register.
3974 @item STAB_REG_TO_REGNUM
3975 @findex STAB_REG_TO_REGNUM
3976 Define this to convert stab register numbers (as gotten from `r'
3977 declarations) into @value{GDBN} regnums. If not defined, no conversion will be
3980 @item DEPRECATED_STACK_ALIGN (@var{addr})
3981 @anchor{DEPRECATED_STACK_ALIGN}
3982 @findex DEPRECATED_STACK_ALIGN
3983 Define this to increase @var{addr} so that it meets the alignment
3984 requirements for the processor's stack.
3986 Unlike @ref{frame_align}, this function always adjusts @var{addr}
3989 By default, no stack alignment is performed.
3991 @item STEP_SKIPS_DELAY (@var{addr})
3992 @findex STEP_SKIPS_DELAY
3993 Define this to return true if the address is of an instruction with a
3994 delay slot. If a breakpoint has been placed in the instruction's delay
3995 slot, @value{GDBN} will single-step over that instruction before resuming
3996 normally. Currently only defined for the Mips.
3998 @item STORE_RETURN_VALUE (@var{type}, @var{regcache}, @var{valbuf})
3999 @findex STORE_RETURN_VALUE
4000 A C expression that writes the function return value, found in
4001 @var{valbuf}, into the @var{regcache}. @var{type} is the type of the
4002 value that is to be returned.
4004 This method has been deprecated in favour of @code{gdbarch_return_value}
4005 (@pxref{gdbarch_return_value}).
4007 @item SYMBOL_RELOADING_DEFAULT
4008 @findex SYMBOL_RELOADING_DEFAULT
4009 The default value of the ``symbol-reloading'' variable. (Never defined in
4012 @item TARGET_CHAR_BIT
4013 @findex TARGET_CHAR_BIT
4014 Number of bits in a char; defaults to 8.
4016 @item TARGET_CHAR_SIGNED
4017 @findex TARGET_CHAR_SIGNED
4018 Non-zero if @code{char} is normally signed on this architecture; zero if
4019 it should be unsigned.
4021 The ISO C standard requires the compiler to treat @code{char} as
4022 equivalent to either @code{signed char} or @code{unsigned char}; any
4023 character in the standard execution set is supposed to be positive.
4024 Most compilers treat @code{char} as signed, but @code{char} is unsigned
4025 on the IBM S/390, RS6000, and PowerPC targets.
4027 @item TARGET_COMPLEX_BIT
4028 @findex TARGET_COMPLEX_BIT
4029 Number of bits in a complex number; defaults to @code{2 * TARGET_FLOAT_BIT}.
4031 At present this macro is not used.
4033 @item TARGET_DOUBLE_BIT
4034 @findex TARGET_DOUBLE_BIT
4035 Number of bits in a double float; defaults to @code{8 * TARGET_CHAR_BIT}.
4037 @item TARGET_DOUBLE_COMPLEX_BIT
4038 @findex TARGET_DOUBLE_COMPLEX_BIT
4039 Number of bits in a double complex; defaults to @code{2 * TARGET_DOUBLE_BIT}.
4041 At present this macro is not used.
4043 @item TARGET_FLOAT_BIT
4044 @findex TARGET_FLOAT_BIT
4045 Number of bits in a float; defaults to @code{4 * TARGET_CHAR_BIT}.
4047 @item TARGET_INT_BIT
4048 @findex TARGET_INT_BIT
4049 Number of bits in an integer; defaults to @code{4 * TARGET_CHAR_BIT}.
4051 @item TARGET_LONG_BIT
4052 @findex TARGET_LONG_BIT
4053 Number of bits in a long integer; defaults to @code{4 * TARGET_CHAR_BIT}.
4055 @item TARGET_LONG_DOUBLE_BIT
4056 @findex TARGET_LONG_DOUBLE_BIT
4057 Number of bits in a long double float;
4058 defaults to @code{2 * TARGET_DOUBLE_BIT}.
4060 @item TARGET_LONG_LONG_BIT
4061 @findex TARGET_LONG_LONG_BIT
4062 Number of bits in a long long integer; defaults to @code{2 * TARGET_LONG_BIT}.
4064 @item TARGET_PTR_BIT
4065 @findex TARGET_PTR_BIT
4066 Number of bits in a pointer; defaults to @code{TARGET_INT_BIT}.
4068 @item TARGET_SHORT_BIT
4069 @findex TARGET_SHORT_BIT
4070 Number of bits in a short integer; defaults to @code{2 * TARGET_CHAR_BIT}.
4072 @item TARGET_READ_PC
4073 @findex TARGET_READ_PC
4074 @itemx TARGET_WRITE_PC (@var{val}, @var{pid})
4075 @findex TARGET_WRITE_PC
4076 @anchor{TARGET_WRITE_PC}
4077 @itemx TARGET_READ_SP
4078 @findex TARGET_READ_SP
4079 @itemx TARGET_READ_FP
4080 @findex TARGET_READ_FP
4085 @anchor{TARGET_READ_SP} These change the behavior of @code{read_pc},
4086 @code{write_pc}, and @code{read_sp}. For most targets, these may be
4087 left undefined. @value{GDBN} will call the read and write register
4088 functions with the relevant @code{_REGNUM} argument.
4090 These macros are useful when a target keeps one of these registers in a
4091 hard to get at place; for example, part in a segment register and part
4092 in an ordinary register.
4094 @xref{unwind_sp}, which replaces @code{TARGET_READ_SP}.
4096 @item TARGET_VIRTUAL_FRAME_POINTER(@var{pc}, @var{regp}, @var{offsetp})
4097 @findex TARGET_VIRTUAL_FRAME_POINTER
4098 Returns a @code{(register, offset)} pair representing the virtual frame
4099 pointer in use at the code address @var{pc}. If virtual frame pointers
4100 are not used, a default definition simply returns
4101 @code{DEPRECATED_FP_REGNUM}, with an offset of zero.
4103 @item TARGET_HAS_HARDWARE_WATCHPOINTS
4104 If non-zero, the target has support for hardware-assisted
4105 watchpoints. @xref{Algorithms, watchpoints}, for more details and
4106 other related macros.
4108 @item TARGET_PRINT_INSN (@var{addr}, @var{info})
4109 @findex TARGET_PRINT_INSN
4110 This is the function used by @value{GDBN} to print an assembly
4111 instruction. It prints the instruction at address @var{addr} in
4112 debugged memory and returns the length of the instruction, in bytes. If
4113 a target doesn't define its own printing routine, it defaults to an
4114 accessor function for the global pointer
4115 @code{deprecated_tm_print_insn}. This usually points to a function in
4116 the @code{opcodes} library (@pxref{Support Libraries, ,Opcodes}).
4117 @var{info} is a structure (of type @code{disassemble_info}) defined in
4118 @file{include/dis-asm.h} used to pass information to the instruction
4121 @item struct frame_id unwind_dummy_id (struct frame_info *@var{frame})
4122 @findex unwind_dummy_id
4123 @anchor{unwind_dummy_id} Given @var{frame} return a @code{struct
4124 frame_id} that uniquely identifies an inferior function call's dummy
4125 frame. The value returned must match the dummy frame stack value
4126 previously saved using @code{SAVE_DUMMY_FRAME_TOS}.
4127 @xref{SAVE_DUMMY_FRAME_TOS}.
4129 @item DEPRECATED_USE_STRUCT_CONVENTION (@var{gcc_p}, @var{type})
4130 @findex DEPRECATED_USE_STRUCT_CONVENTION
4131 If defined, this must be an expression that is nonzero if a value of the
4132 given @var{type} being returned from a function must have space
4133 allocated for it on the stack. @var{gcc_p} is true if the function
4134 being considered is known to have been compiled by GCC; this is helpful
4135 for systems where GCC is known to use different calling convention than
4138 This method has been deprecated in favour of @code{gdbarch_return_value}
4139 (@pxref{gdbarch_return_value}).
4141 @item VALUE_TO_REGISTER(@var{type}, @var{regnum}, @var{from}, @var{to})
4142 @findex VALUE_TO_REGISTER
4143 Convert a value of type @var{type} into the raw contents of register
4145 @xref{Target Architecture Definition, , Using Different Register and Memory Data Representations}.
4147 @item VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
4148 @findex VARIABLES_INSIDE_BLOCK
4149 For dbx-style debugging information, if the compiler puts variable
4150 declarations inside LBRAC/RBRAC blocks, this should be defined to be
4151 nonzero. @var{desc} is the value of @code{n_desc} from the
4152 @code{N_RBRAC} symbol, and @var{gcc_p} is true if @value{GDBN} has noticed the
4153 presence of either the @code{GCC_COMPILED_SYMBOL} or the
4154 @code{GCC2_COMPILED_SYMBOL}. By default, this is 0.
4156 @item OS9K_VARIABLES_INSIDE_BLOCK (@var{desc}, @var{gcc_p})
4157 @findex OS9K_VARIABLES_INSIDE_BLOCK
4158 Similarly, for OS/9000. Defaults to 1.
4161 Motorola M68K target conditionals.
4165 Define this to be the 4-bit location of the breakpoint trap vector. If
4166 not defined, it will default to @code{0xf}.
4168 @item REMOTE_BPT_VECTOR
4169 Defaults to @code{1}.
4171 @item NAME_OF_MALLOC
4172 @findex NAME_OF_MALLOC
4173 A string containing the name of the function to call in order to
4174 allocate some memory in the inferior. The default value is "malloc".
4178 @section Adding a New Target
4180 @cindex adding a target
4181 The following files add a target to @value{GDBN}:
4185 @item gdb/config/@var{arch}/@var{ttt}.mt
4186 Contains a Makefile fragment specific to this target. Specifies what
4187 object files are needed for target @var{ttt}, by defining
4188 @samp{TDEPFILES=@dots{}} and @samp{TDEPLIBS=@dots{}}. Also specifies
4189 the header file which describes @var{ttt}, by defining @samp{TM_FILE=
4192 You can also define @samp{TM_CFLAGS}, @samp{TM_CLIBS}, @samp{TM_CDEPS},
4193 but these are now deprecated, replaced by autoconf, and may go away in
4194 future versions of @value{GDBN}.
4196 @item gdb/@var{ttt}-tdep.c
4197 Contains any miscellaneous code required for this target machine. On
4198 some machines it doesn't exist at all. Sometimes the macros in
4199 @file{tm-@var{ttt}.h} become very complicated, so they are implemented
4200 as functions here instead, and the macro is simply defined to call the
4201 function. This is vastly preferable, since it is easier to understand
4204 @item gdb/@var{arch}-tdep.c
4205 @itemx gdb/@var{arch}-tdep.h
4206 This often exists to describe the basic layout of the target machine's
4207 processor chip (registers, stack, etc.). If used, it is included by
4208 @file{@var{ttt}-tdep.h}. It can be shared among many targets that use
4211 @item gdb/config/@var{arch}/tm-@var{ttt}.h
4212 (@file{tm.h} is a link to this file, created by @code{configure}). Contains
4213 macro definitions about the target machine's registers, stack frame
4214 format and instructions.
4216 New targets do not need this file and should not create it.
4218 @item gdb/config/@var{arch}/tm-@var{arch}.h
4219 This often exists to describe the basic layout of the target machine's
4220 processor chip (registers, stack, etc.). If used, it is included by
4221 @file{tm-@var{ttt}.h}. It can be shared among many targets that use the
4224 New targets do not need this file and should not create it.
4228 If you are adding a new operating system for an existing CPU chip, add a
4229 @file{config/tm-@var{os}.h} file that describes the operating system
4230 facilities that are unusual (extra symbol table info; the breakpoint
4231 instruction needed; etc.). Then write a @file{@var{arch}/tm-@var{os}.h}
4232 that just @code{#include}s @file{tm-@var{arch}.h} and
4233 @file{config/tm-@var{os}.h}.
4236 @section Converting an existing Target Architecture to Multi-arch
4237 @cindex converting targets to multi-arch
4239 This section describes the current accepted best practice for converting
4240 an existing target architecture to the multi-arch framework.
4242 The process consists of generating, testing, posting and committing a
4243 sequence of patches. Each patch must contain a single change, for
4249 Directly convert a group of functions into macros (the conversion does
4250 not change the behavior of any of the functions).
4253 Replace a non-multi-arch with a multi-arch mechanism (e.g.,
4257 Enable multi-arch level one.
4260 Delete one or more files.
4265 There isn't a size limit on a patch, however, a developer is strongly
4266 encouraged to keep the patch size down.
4268 Since each patch is well defined, and since each change has been tested
4269 and shows no regressions, the patches are considered @emph{fairly}
4270 obvious. Such patches, when submitted by developers listed in the
4271 @file{MAINTAINERS} file, do not need approval. Occasional steps in the
4272 process may be more complicated and less clear. The developer is
4273 expected to use their judgment and is encouraged to seek advice as
4276 @subsection Preparation
4278 The first step is to establish control. Build (with @option{-Werror}
4279 enabled) and test the target so that there is a baseline against which
4280 the debugger can be compared.
4282 At no stage can the test results regress or @value{GDBN} stop compiling
4283 with @option{-Werror}.
4285 @subsection Add the multi-arch initialization code
4287 The objective of this step is to establish the basic multi-arch
4288 framework. It involves
4293 The addition of a @code{@var{arch}_gdbarch_init} function@footnote{The
4294 above is from the original example and uses K&R C. @value{GDBN}
4295 has since converted to ISO C but lets ignore that.} that creates
4298 static struct gdbarch *
4299 d10v_gdbarch_init (info, arches)
4300 struct gdbarch_info info;
4301 struct gdbarch_list *arches;
4303 struct gdbarch *gdbarch;
4304 /* there is only one d10v architecture */
4306 return arches->gdbarch;
4307 gdbarch = gdbarch_alloc (&info, NULL);
4315 A per-architecture dump function to print any architecture specific
4319 mips_dump_tdep (struct gdbarch *current_gdbarch,
4320 struct ui_file *file)
4322 @dots{} code to print architecture specific info @dots{}
4327 A change to @code{_initialize_@var{arch}_tdep} to register this new
4331 _initialize_mips_tdep (void)
4333 gdbarch_register (bfd_arch_mips, mips_gdbarch_init,
4338 Add the macro @code{GDB_MULTI_ARCH}, defined as 0 (zero), to the file@*
4339 @file{config/@var{arch}/tm-@var{arch}.h}.
4343 @subsection Update multi-arch incompatible mechanisms
4345 Some mechanisms do not work with multi-arch. They include:
4348 @item FRAME_FIND_SAVED_REGS
4349 Replaced with @code{DEPRECATED_FRAME_INIT_SAVED_REGS}
4353 At this stage you could also consider converting the macros into
4356 @subsection Prepare for multi-arch level to one
4358 Temporally set @code{GDB_MULTI_ARCH} to @code{GDB_MULTI_ARCH_PARTIAL}
4359 and then build and start @value{GDBN} (the change should not be
4360 committed). @value{GDBN} may not build, and once built, it may die with
4361 an internal error listing the architecture methods that must be
4364 Fix any build problems (patch(es)).
4366 Convert all the architecture methods listed, which are only macros, into
4367 functions (patch(es)).
4369 Update @code{@var{arch}_gdbarch_init} to set all the missing
4370 architecture methods and wrap the corresponding macros in @code{#if
4371 !GDB_MULTI_ARCH} (patch(es)).
4373 @subsection Set multi-arch level one
4375 Change the value of @code{GDB_MULTI_ARCH} to GDB_MULTI_ARCH_PARTIAL (a
4378 Any problems with throwing ``the switch'' should have been fixed
4381 @subsection Convert remaining macros
4383 Suggest converting macros into functions (and setting the corresponding
4384 architecture method) in small batches.
4386 @subsection Set multi-arch level to two
4388 This should go smoothly.
4390 @subsection Delete the TM file
4392 The @file{tm-@var{arch}.h} can be deleted. @file{@var{arch}.mt} and
4393 @file{configure.in} updated.
4396 @node Target Vector Definition
4398 @chapter Target Vector Definition
4399 @cindex target vector
4401 The target vector defines the interface between @value{GDBN}'s
4402 abstract handling of target systems, and the nitty-gritty code that
4403 actually exercises control over a process or a serial port.
4404 @value{GDBN} includes some 30-40 different target vectors; however,
4405 each configuration of @value{GDBN} includes only a few of them.
4407 @section File Targets
4409 Both executables and core files have target vectors.
4411 @section Standard Protocol and Remote Stubs
4413 @value{GDBN}'s file @file{remote.c} talks a serial protocol to code
4414 that runs in the target system. @value{GDBN} provides several sample
4415 @dfn{stubs} that can be integrated into target programs or operating
4416 systems for this purpose; they are named @file{*-stub.c}.
4418 The @value{GDBN} user's manual describes how to put such a stub into
4419 your target code. What follows is a discussion of integrating the
4420 SPARC stub into a complicated operating system (rather than a simple
4421 program), by Stu Grossman, the author of this stub.
4423 The trap handling code in the stub assumes the following upon entry to
4428 %l1 and %l2 contain pc and npc respectively at the time of the trap;
4434 you are in the correct trap window.
4437 As long as your trap handler can guarantee those conditions, then there
4438 is no reason why you shouldn't be able to ``share'' traps with the stub.
4439 The stub has no requirement that it be jumped to directly from the
4440 hardware trap vector. That is why it calls @code{exceptionHandler()},
4441 which is provided by the external environment. For instance, this could
4442 set up the hardware traps to actually execute code which calls the stub
4443 first, and then transfers to its own trap handler.
4445 For the most point, there probably won't be much of an issue with
4446 ``sharing'' traps, as the traps we use are usually not used by the kernel,
4447 and often indicate unrecoverable error conditions. Anyway, this is all
4448 controlled by a table, and is trivial to modify. The most important
4449 trap for us is for @code{ta 1}. Without that, we can't single step or
4450 do breakpoints. Everything else is unnecessary for the proper operation
4451 of the debugger/stub.
4453 From reading the stub, it's probably not obvious how breakpoints work.
4454 They are simply done by deposit/examine operations from @value{GDBN}.
4456 @section ROM Monitor Interface
4458 @section Custom Protocols
4460 @section Transport Layer
4462 @section Builtin Simulator
4465 @node Native Debugging
4467 @chapter Native Debugging
4468 @cindex native debugging
4470 Several files control @value{GDBN}'s configuration for native support:
4474 @item gdb/config/@var{arch}/@var{xyz}.mh
4475 Specifies Makefile fragments needed by a @emph{native} configuration on
4476 machine @var{xyz}. In particular, this lists the required
4477 native-dependent object files, by defining @samp{NATDEPFILES=@dots{}}.
4478 Also specifies the header file which describes native support on
4479 @var{xyz}, by defining @samp{NAT_FILE= nm-@var{xyz}.h}. You can also
4480 define @samp{NAT_CFLAGS}, @samp{NAT_ADD_FILES}, @samp{NAT_CLIBS},
4481 @samp{NAT_CDEPS}, etc.; see @file{Makefile.in}.
4483 @emph{Maintainer's note: The @file{.mh} suffix is because this file
4484 originally contained @file{Makefile} fragments for hosting @value{GDBN}
4485 on machine @var{xyz}. While the file is no longer used for this
4486 purpose, the @file{.mh} suffix remains. Perhaps someone will
4487 eventually rename these fragments so that they have a @file{.mn}
4490 @item gdb/config/@var{arch}/nm-@var{xyz}.h
4491 (@file{nm.h} is a link to this file, created by @code{configure}). Contains C
4492 macro definitions describing the native system environment, such as
4493 child process control and core file support.
4495 @item gdb/@var{xyz}-nat.c
4496 Contains any miscellaneous C code required for this native support of
4497 this machine. On some machines it doesn't exist at all.
4500 There are some ``generic'' versions of routines that can be used by
4501 various systems. These can be customized in various ways by macros
4502 defined in your @file{nm-@var{xyz}.h} file. If these routines work for
4503 the @var{xyz} host, you can just include the generic file's name (with
4504 @samp{.o}, not @samp{.c}) in @code{NATDEPFILES}.
4506 Otherwise, if your machine needs custom support routines, you will need
4507 to write routines that perform the same functions as the generic file.
4508 Put them into @file{@var{xyz}-nat.c}, and put @file{@var{xyz}-nat.o}
4509 into @code{NATDEPFILES}.
4513 This contains the @emph{target_ops vector} that supports Unix child
4514 processes on systems which use ptrace and wait to control the child.
4517 This contains the @emph{target_ops vector} that supports Unix child
4518 processes on systems which use /proc to control the child.
4521 This does the low-level grunge that uses Unix system calls to do a ``fork
4522 and exec'' to start up a child process.
4525 This is the low level interface to inferior processes for systems using
4526 the Unix @code{ptrace} call in a vanilla way.
4529 @section Native core file Support
4530 @cindex native core files
4533 @findex fetch_core_registers
4534 @item core-aout.c::fetch_core_registers()
4535 Support for reading registers out of a core file. This routine calls
4536 @code{register_addr()}, see below. Now that BFD is used to read core
4537 files, virtually all machines should use @code{core-aout.c}, and should
4538 just provide @code{fetch_core_registers} in @code{@var{xyz}-nat.c} (or
4539 @code{REGISTER_U_ADDR} in @code{nm-@var{xyz}.h}).
4541 @item core-aout.c::register_addr()
4542 If your @code{nm-@var{xyz}.h} file defines the macro
4543 @code{REGISTER_U_ADDR(addr, blockend, regno)}, it should be defined to
4544 set @code{addr} to the offset within the @samp{user} struct of @value{GDBN}
4545 register number @code{regno}. @code{blockend} is the offset within the
4546 ``upage'' of @code{u.u_ar0}. If @code{REGISTER_U_ADDR} is defined,
4547 @file{core-aout.c} will define the @code{register_addr()} function and
4548 use the macro in it. If you do not define @code{REGISTER_U_ADDR}, but
4549 you are using the standard @code{fetch_core_registers()}, you will need
4550 to define your own version of @code{register_addr()}, put it into your
4551 @code{@var{xyz}-nat.c} file, and be sure @code{@var{xyz}-nat.o} is in
4552 the @code{NATDEPFILES} list. If you have your own
4553 @code{fetch_core_registers()}, you may not need a separate
4554 @code{register_addr()}. Many custom @code{fetch_core_registers()}
4555 implementations simply locate the registers themselves.@refill
4558 When making @value{GDBN} run native on a new operating system, to make it
4559 possible to debug core files, you will need to either write specific
4560 code for parsing your OS's core files, or customize
4561 @file{bfd/trad-core.c}. First, use whatever @code{#include} files your
4562 machine uses to define the struct of registers that is accessible
4563 (possibly in the u-area) in a core file (rather than
4564 @file{machine/reg.h}), and an include file that defines whatever header
4565 exists on a core file (e.g., the u-area or a @code{struct core}). Then
4566 modify @code{trad_unix_core_file_p} to use these values to set up the
4567 section information for the data segment, stack segment, any other
4568 segments in the core file (perhaps shared library contents or control
4569 information), ``registers'' segment, and if there are two discontiguous
4570 sets of registers (e.g., integer and float), the ``reg2'' segment. This
4571 section information basically delimits areas in the core file in a
4572 standard way, which the section-reading routines in BFD know how to seek
4575 Then back in @value{GDBN}, you need a matching routine called
4576 @code{fetch_core_registers}. If you can use the generic one, it's in
4577 @file{core-aout.c}; if not, it's in your @file{@var{xyz}-nat.c} file.
4578 It will be passed a char pointer to the entire ``registers'' segment,
4579 its length, and a zero; or a char pointer to the entire ``regs2''
4580 segment, its length, and a 2. The routine should suck out the supplied
4581 register values and install them into @value{GDBN}'s ``registers'' array.
4583 If your system uses @file{/proc} to control processes, and uses ELF
4584 format core files, then you may be able to use the same routines for
4585 reading the registers out of processes and out of core files.
4593 @section shared libraries
4595 @section Native Conditionals
4596 @cindex native conditionals
4598 When @value{GDBN} is configured and compiled, various macros are
4599 defined or left undefined, to control compilation when the host and
4600 target systems are the same. These macros should be defined (or left
4601 undefined) in @file{nm-@var{system}.h}.
4605 @item CHILD_PREPARE_TO_STORE
4606 @findex CHILD_PREPARE_TO_STORE
4607 If the machine stores all registers at once in the child process, then
4608 define this to ensure that all values are correct. This usually entails
4609 a read from the child.
4611 [Note that this is incorrectly defined in @file{xm-@var{system}.h} files
4614 @item FETCH_INFERIOR_REGISTERS
4615 @findex FETCH_INFERIOR_REGISTERS
4616 Define this if the native-dependent code will provide its own routines
4617 @code{fetch_inferior_registers} and @code{store_inferior_registers} in
4618 @file{@var{host}-nat.c}. If this symbol is @emph{not} defined, and
4619 @file{infptrace.c} is included in this configuration, the default
4620 routines in @file{infptrace.c} are used for these functions.
4624 This macro is normally defined to be the number of the first floating
4625 point register, if the machine has such registers. As such, it would
4626 appear only in target-specific code. However, @file{/proc} support uses this
4627 to decide whether floats are in use on this target.
4629 @item GET_LONGJMP_TARGET
4630 @findex GET_LONGJMP_TARGET
4631 For most machines, this is a target-dependent parameter. On the
4632 DECstation and the Iris, this is a native-dependent parameter, since
4633 @file{setjmp.h} is needed to define it.
4635 This macro determines the target PC address that @code{longjmp} will jump to,
4636 assuming that we have just stopped at a longjmp breakpoint. It takes a
4637 @code{CORE_ADDR *} as argument, and stores the target PC value through this
4638 pointer. It examines the current state of the machine as needed.
4640 @item I386_USE_GENERIC_WATCHPOINTS
4641 An x86-based machine can define this to use the generic x86 watchpoint
4642 support; see @ref{Algorithms, I386_USE_GENERIC_WATCHPOINTS}.
4645 @findex KERNEL_U_ADDR
4646 Define this to the address of the @code{u} structure (the ``user
4647 struct'', also known as the ``u-page'') in kernel virtual memory. @value{GDBN}
4648 needs to know this so that it can subtract this address from absolute
4649 addresses in the upage, that are obtained via ptrace or from core files.
4650 On systems that don't need this value, set it to zero.
4652 @item KERNEL_U_ADDR_HPUX
4653 @findex KERNEL_U_ADDR_HPUX
4654 Define this to cause @value{GDBN} to determine the address of @code{u} at
4655 runtime, by using HP-style @code{nlist} on the kernel's image in the
4658 @item ONE_PROCESS_WRITETEXT
4659 @findex ONE_PROCESS_WRITETEXT
4660 Define this to be able to, when a breakpoint insertion fails, warn the
4661 user that another process may be running with the same executable.
4664 @findex PROC_NAME_FMT
4665 Defines the format for the name of a @file{/proc} device. Should be
4666 defined in @file{nm.h} @emph{only} in order to override the default
4667 definition in @file{procfs.c}.
4669 @item PTRACE_ARG3_TYPE
4670 @findex PTRACE_ARG3_TYPE
4671 The type of the third argument to the @code{ptrace} system call, if it
4672 exists and is different from @code{int}.
4674 @item REGISTER_U_ADDR
4675 @findex REGISTER_U_ADDR
4676 Defines the offset of the registers in the ``u area''.
4678 @item SHELL_COMMAND_CONCAT
4679 @findex SHELL_COMMAND_CONCAT
4680 If defined, is a string to prefix on the shell command used to start the
4685 If defined, this is the name of the shell to use to run the inferior.
4686 Defaults to @code{"/bin/sh"}.
4688 @item SOLIB_ADD (@var{filename}, @var{from_tty}, @var{targ}, @var{readsyms})
4690 Define this to expand into an expression that will cause the symbols in
4691 @var{filename} to be added to @value{GDBN}'s symbol table. If
4692 @var{readsyms} is zero symbols are not read but any necessary low level
4693 processing for @var{filename} is still done.
4695 @item SOLIB_CREATE_INFERIOR_HOOK
4696 @findex SOLIB_CREATE_INFERIOR_HOOK
4697 Define this to expand into any shared-library-relocation code that you
4698 want to be run just after the child process has been forked.
4700 @item START_INFERIOR_TRAPS_EXPECTED
4701 @findex START_INFERIOR_TRAPS_EXPECTED
4702 When starting an inferior, @value{GDBN} normally expects to trap
4704 the shell execs, and once when the program itself execs. If the actual
4705 number of traps is something other than 2, then define this macro to
4706 expand into the number expected.
4710 This determines whether small routines in @file{*-tdep.c}, which
4711 translate register values between @value{GDBN}'s internal
4712 representation and the @file{/proc} representation, are compiled.
4715 @findex U_REGS_OFFSET
4716 This is the offset of the registers in the upage. It need only be
4717 defined if the generic ptrace register access routines in
4718 @file{infptrace.c} are being used (that is, @file{infptrace.c} is
4719 configured in, and @code{FETCH_INFERIOR_REGISTERS} is not defined). If
4720 the default value from @file{infptrace.c} is good enough, leave it
4723 The default value means that u.u_ar0 @emph{points to} the location of
4724 the registers. I'm guessing that @code{#define U_REGS_OFFSET 0} means
4725 that @code{u.u_ar0} @emph{is} the location of the registers.
4729 See @file{objfiles.c}.
4732 @findex DEBUG_PTRACE
4733 Define this to debug @code{ptrace} calls.
4737 @node Support Libraries
4739 @chapter Support Libraries
4744 BFD provides support for @value{GDBN} in several ways:
4747 @item identifying executable and core files
4748 BFD will identify a variety of file types, including a.out, coff, and
4749 several variants thereof, as well as several kinds of core files.
4751 @item access to sections of files
4752 BFD parses the file headers to determine the names, virtual addresses,
4753 sizes, and file locations of all the various named sections in files
4754 (such as the text section or the data section). @value{GDBN} simply
4755 calls BFD to read or write section @var{x} at byte offset @var{y} for
4758 @item specialized core file support
4759 BFD provides routines to determine the failing command name stored in a
4760 core file, the signal with which the program failed, and whether a core
4761 file matches (i.e.@: could be a core dump of) a particular executable
4764 @item locating the symbol information
4765 @value{GDBN} uses an internal interface of BFD to determine where to find the
4766 symbol information in an executable file or symbol-file. @value{GDBN} itself
4767 handles the reading of symbols, since BFD does not ``understand'' debug
4768 symbols, but @value{GDBN} uses BFD's cached information to find the symbols,
4773 @cindex opcodes library
4775 The opcodes library provides @value{GDBN}'s disassembler. (It's a separate
4776 library because it's also used in binutils, for @file{objdump}).
4783 @cindex @code{libiberty} library
4785 The @code{libiberty} library provides a set of functions and features
4786 that integrate and improve on functionality found in modern operating
4787 systems. Broadly speaking, such features can be divided into three
4788 groups: supplemental functions (functions that may be missing in some
4789 environments and operating systems), replacement functions (providing
4790 a uniform and easier to use interface for commonly used standard
4791 functions), and extensions (which provide additional functionality
4792 beyond standard functions).
4794 @value{GDBN} uses various features provided by the @code{libiberty}
4795 library, for instance the C@t{++} demangler, the @acronym{IEEE}
4796 floating format support functions, the input options parser
4797 @samp{getopt}, the @samp{obstack} extension, and other functions.
4799 @subsection @code{obstacks} in @value{GDBN}
4800 @cindex @code{obstacks}
4802 The obstack mechanism provides a convenient way to allocate and free
4803 chunks of memory. Each obstack is a pool of memory that is managed
4804 like a stack. Objects (of any nature, size and alignment) are
4805 allocated and freed in a @acronym{LIFO} fashion on an obstack (see
4806 @code{libiberty}'s documenatation for a more detailed explanation of
4809 The most noticeable use of the @code{obstacks} in @value{GDBN} is in
4810 object files. There is an obstack associated with each internal
4811 representation of an object file. Lots of things get allocated on
4812 these @code{obstacks}: dictionary entries, blocks, blockvectors,
4813 symbols, minimal symbols, types, vectors of fundamental types, class
4814 fields of types, object files section lists, object files section
4815 offets lists, line tables, symbol tables, partial symbol tables,
4816 string tables, symbol table private data, macros tables, debug
4817 information sections and entries, import and export lists (som),
4818 unwind information (hppa), dwarf2 location expressions data. Plus
4819 various strings such as directory names strings, debug format strings,
4822 An essential and convenient property of all data on @code{obstacks} is
4823 that memory for it gets allocated (with @code{obstack_alloc}) at
4824 various times during a debugging sesssion, but it is released all at
4825 once using the @code{obstack_free} function. The @code{obstack_free}
4826 function takes a pointer to where in the stack it must start the
4827 deletion from (much like the cleanup chains have a pointer to where to
4828 start the cleanups). Because of the stack like structure of the
4829 @code{obstacks}, this allows to free only a top portion of the
4830 obstack. There are a few instances in @value{GDBN} where such thing
4831 happens. Calls to @code{obstack_free} are done after some local data
4832 is allocated to the obstack. Only the local data is deleted from the
4833 obstack. Of course this assumes that nothing between the
4834 @code{obstack_alloc} and the @code{obstack_free} allocates anything
4835 else on the same obstack. For this reason it is best and safest to
4836 use temporary @code{obstacks}.
4838 Releasing the whole obstack is also not safe per se. It is safe only
4839 under the condition that we know the @code{obstacks} memory is no
4840 longer needed. In @value{GDBN} we get rid of the @code{obstacks} only
4841 when we get rid of the whole objfile(s), for instance upon reading a
4845 @cindex regular expressions library
4856 @item SIGN_EXTEND_CHAR
4858 @item SWITCH_ENUM_BUG
4873 This chapter covers topics that are lower-level than the major
4874 algorithms of @value{GDBN}.
4879 Cleanups are a structured way to deal with things that need to be done
4882 When your code does something (e.g., @code{xmalloc} some memory, or
4883 @code{open} a file) that needs to be undone later (e.g., @code{xfree}
4884 the memory or @code{close} the file), it can make a cleanup. The
4885 cleanup will be done at some future point: when the command is finished
4886 and control returns to the top level; when an error occurs and the stack
4887 is unwound; or when your code decides it's time to explicitly perform
4888 cleanups. Alternatively you can elect to discard the cleanups you
4894 @item struct cleanup *@var{old_chain};
4895 Declare a variable which will hold a cleanup chain handle.
4897 @findex make_cleanup
4898 @item @var{old_chain} = make_cleanup (@var{function}, @var{arg});
4899 Make a cleanup which will cause @var{function} to be called with
4900 @var{arg} (a @code{char *}) later. The result, @var{old_chain}, is a
4901 handle that can later be passed to @code{do_cleanups} or
4902 @code{discard_cleanups}. Unless you are going to call
4903 @code{do_cleanups} or @code{discard_cleanups}, you can ignore the result
4904 from @code{make_cleanup}.
4907 @item do_cleanups (@var{old_chain});
4908 Do all cleanups added to the chain since the corresponding
4909 @code{make_cleanup} call was made.
4911 @findex discard_cleanups
4912 @item discard_cleanups (@var{old_chain});
4913 Same as @code{do_cleanups} except that it just removes the cleanups from
4914 the chain and does not call the specified functions.
4917 Cleanups are implemented as a chain. The handle returned by
4918 @code{make_cleanups} includes the cleanup passed to the call and any
4919 later cleanups appended to the chain (but not yet discarded or
4923 make_cleanup (a, 0);
4925 struct cleanup *old = make_cleanup (b, 0);
4933 will call @code{c()} and @code{b()} but will not call @code{a()}. The
4934 cleanup that calls @code{a()} will remain in the cleanup chain, and will
4935 be done later unless otherwise discarded.@refill
4937 Your function should explicitly do or discard the cleanups it creates.
4938 Failing to do this leads to non-deterministic behavior since the caller
4939 will arbitrarily do or discard your functions cleanups. This need leads
4940 to two common cleanup styles.
4942 The first style is try/finally. Before it exits, your code-block calls
4943 @code{do_cleanups} with the old cleanup chain and thus ensures that your
4944 code-block's cleanups are always performed. For instance, the following
4945 code-segment avoids a memory leak problem (even when @code{error} is
4946 called and a forced stack unwind occurs) by ensuring that the
4947 @code{xfree} will always be called:
4950 struct cleanup *old = make_cleanup (null_cleanup, 0);
4951 data = xmalloc (sizeof blah);
4952 make_cleanup (xfree, data);
4957 The second style is try/except. Before it exits, your code-block calls
4958 @code{discard_cleanups} with the old cleanup chain and thus ensures that
4959 any created cleanups are not performed. For instance, the following
4960 code segment, ensures that the file will be closed but only if there is
4964 FILE *file = fopen ("afile", "r");
4965 struct cleanup *old = make_cleanup (close_file, file);
4967 discard_cleanups (old);
4971 Some functions, e.g., @code{fputs_filtered()} or @code{error()}, specify
4972 that they ``should not be called when cleanups are not in place''. This
4973 means that any actions you need to reverse in the case of an error or
4974 interruption must be on the cleanup chain before you call these
4975 functions, since they might never return to your code (they
4976 @samp{longjmp} instead).
4978 @section Per-architecture module data
4979 @cindex per-architecture module data
4980 @cindex multi-arch data
4981 @cindex data-pointer, per-architecture/per-module
4983 The multi-arch framework includes a mechanism for adding module
4984 specific per-architecture data-pointers to the @code{struct gdbarch}
4985 architecture object.
4987 A module registers one or more per-architecture data-pointers using:
4989 @deftypefun struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *@var{pre_init})
4990 @var{pre_init} is used to, on-demand, allocate an initial value for a
4991 per-architecture data-pointer using the architecture's obstack (passed
4992 in as a parameter). Since @var{pre_init} can be called during
4993 architecture creation, it is not parameterized with the architecture.
4994 and must not call modules that use per-architecture data.
4997 @deftypefun struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *@var{post_init})
4998 @var{post_init} is used to obtain an initial value for a
4999 per-architecture data-pointer @emph{after}. Since @var{post_init} is
5000 always called after architecture creation, it both receives the fully
5001 initialized architecture and is free to call modules that use
5002 per-architecture data (care needs to be taken to ensure that those
5003 other modules do not try to call back to this module as that will
5004 create in cycles in the initialization call graph).
5007 These functions return a @code{struct gdbarch_data} that is used to
5008 identify the per-architecture data-pointer added for that module.
5010 The per-architecture data-pointer is accessed using the function:
5012 @deftypefun void *gdbarch_data (struct gdbarch *@var{gdbarch}, struct gdbarch_data *@var{data_handle})
5013 Given the architecture @var{arch} and module data handle
5014 @var{data_handle} (returned by @code{gdbarch_data_register_pre_init}
5015 or @code{gdbarch_data_register_post_init}), this function returns the
5016 current value of the per-architecture data-pointer. If the data
5017 pointer is @code{NULL}, it is first initialized by calling the
5018 corresponding @var{pre_init} or @var{post_init} method.
5021 The examples below assume the following definitions:
5024 struct nozel @{ int total; @};
5025 static struct gdbarch_data *nozel_handle;
5028 A module can extend the architecture vector, adding additional
5029 per-architecture data, using the @var{pre_init} method. The module's
5030 per-architecture data is then initialized during architecture
5033 In the below, the module's per-architecture @emph{nozel} is added. An
5034 architecture can specify its nozel by calling @code{set_gdbarch_nozel}
5035 from @code{gdbarch_init}.
5039 nozel_pre_init (struct obstack *obstack)
5041 struct nozel *data = OBSTACK_ZALLOC (obstack, struct nozel);
5048 set_gdbarch_nozel (struct gdbarch *gdbarch, int total)
5050 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5051 data->total = nozel;
5055 A module can on-demand create architecture dependant data structures
5056 using @code{post_init}.
5058 In the below, the nozel's total is computed on-demand by
5059 @code{nozel_post_init} using information obtained from the
5064 nozel_post_init (struct gdbarch *gdbarch)
5066 struct nozel *data = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct nozel);
5067 nozel->total = gdbarch@dots{} (gdbarch);
5074 nozel_total (struct gdbarch *gdbarch)
5076 struct nozel *data = gdbarch_data (gdbarch, nozel_handle);
5081 @section Wrapping Output Lines
5082 @cindex line wrap in output
5085 Output that goes through @code{printf_filtered} or @code{fputs_filtered}
5086 or @code{fputs_demangled} needs only to have calls to @code{wrap_here}
5087 added in places that would be good breaking points. The utility
5088 routines will take care of actually wrapping if the line width is
5091 The argument to @code{wrap_here} is an indentation string which is
5092 printed @emph{only} if the line breaks there. This argument is saved
5093 away and used later. It must remain valid until the next call to
5094 @code{wrap_here} or until a newline has been printed through the
5095 @code{*_filtered} functions. Don't pass in a local variable and then
5098 It is usually best to call @code{wrap_here} after printing a comma or
5099 space. If you call it before printing a space, make sure that your
5100 indentation properly accounts for the leading space that will print if
5101 the line wraps there.
5103 Any function or set of functions that produce filtered output must
5104 finish by printing a newline, to flush the wrap buffer, before switching
5105 to unfiltered (@code{printf}) output. Symbol reading routines that
5106 print warnings are a good example.
5108 @section @value{GDBN} Coding Standards
5109 @cindex coding standards
5111 @value{GDBN} follows the GNU coding standards, as described in
5112 @file{etc/standards.texi}. This file is also available for anonymous
5113 FTP from GNU archive sites. @value{GDBN} takes a strict interpretation
5114 of the standard; in general, when the GNU standard recommends a practice
5115 but does not require it, @value{GDBN} requires it.
5117 @value{GDBN} follows an additional set of coding standards specific to
5118 @value{GDBN}, as described in the following sections.
5123 @value{GDBN} assumes an ISO/IEC 9899:1990 (a.k.a.@: ISO C90) compliant
5126 @value{GDBN} does not assume an ISO C or POSIX compliant C library.
5129 @subsection Memory Management
5131 @value{GDBN} does not use the functions @code{malloc}, @code{realloc},
5132 @code{calloc}, @code{free} and @code{asprintf}.
5134 @value{GDBN} uses the functions @code{xmalloc}, @code{xrealloc} and
5135 @code{xcalloc} when allocating memory. Unlike @code{malloc} et.al.@:
5136 these functions do not return when the memory pool is empty. Instead,
5137 they unwind the stack using cleanups. These functions return
5138 @code{NULL} when requested to allocate a chunk of memory of size zero.
5140 @emph{Pragmatics: By using these functions, the need to check every
5141 memory allocation is removed. These functions provide portable
5144 @value{GDBN} does not use the function @code{free}.
5146 @value{GDBN} uses the function @code{xfree} to return memory to the
5147 memory pool. Consistent with ISO-C, this function ignores a request to
5148 free a @code{NULL} pointer.
5150 @emph{Pragmatics: On some systems @code{free} fails when passed a
5151 @code{NULL} pointer.}
5153 @value{GDBN} can use the non-portable function @code{alloca} for the
5154 allocation of small temporary values (such as strings).
5156 @emph{Pragmatics: This function is very non-portable. Some systems
5157 restrict the memory being allocated to no more than a few kilobytes.}
5159 @value{GDBN} uses the string function @code{xstrdup} and the print
5160 function @code{xstrprintf}.
5162 @emph{Pragmatics: @code{asprintf} and @code{strdup} can fail. Print
5163 functions such as @code{sprintf} are very prone to buffer overflow
5167 @subsection Compiler Warnings
5168 @cindex compiler warnings
5170 With few exceptions, developers should include the configuration option
5171 @samp{--enable-gdb-build-warnings=,-Werror} when building @value{GDBN}.
5172 The exceptions are listed in the file @file{gdb/MAINTAINERS}.
5174 This option causes @value{GDBN} (when built using GCC) to be compiled
5175 with a carefully selected list of compiler warning flags. Any warnings
5176 from those flags being treated as errors.
5178 The current list of warning flags includes:
5182 Since @value{GDBN} coding standard requires all functions to be declared
5183 using a prototype, the flag has the side effect of ensuring that
5184 prototyped functions are always visible with out resorting to
5185 @samp{-Wstrict-prototypes}.
5188 Such code often appears to work except on instruction set architectures
5189 that use register windows.
5196 @itemx -Wformat-nonliteral
5197 Since @value{GDBN} uses the @code{format printf} attribute on all
5198 @code{printf} like functions these check not just @code{printf} calls
5199 but also calls to functions such as @code{fprintf_unfiltered}.
5202 This warning includes uses of the assignment operator within an
5203 @code{if} statement.
5205 @item -Wpointer-arith
5207 @item -Wuninitialized
5209 @item -Wunused-label
5210 This warning has the additional benefit of detecting the absence of the
5211 @code{case} reserved word in a switch statement:
5213 enum @{ FD_SCHEDULED, NOTHING_SCHEDULED @} sched;
5226 @item -Wunused-function
5228 @item -Wno-pointer-sign
5229 In version 4.0, GCC began warning about pointer argument passing or
5230 assignment even when the source and destination differed only in
5231 signedness. However, most @value{GDBN} code doesn't distinguish
5232 carefully between @code{char} and @code{unsigned char}. In early 2006
5233 the @value{GDBN} developers decided correcting these warnings wasn't
5234 worth the time it would take.
5238 @emph{Pragmatics: Due to the way that @value{GDBN} is implemented most
5239 functions have unused parameters. Consequently the warning
5240 @samp{-Wunused-parameter} is precluded from the list. The macro
5241 @code{ATTRIBUTE_UNUSED} is not used as it leads to false negatives ---
5242 it is not an error to have @code{ATTRIBUTE_UNUSED} on a parameter that
5243 is being used. The options @samp{-Wall} and @samp{-Wunused} are also
5244 precluded because they both include @samp{-Wunused-parameter}.}
5246 @emph{Pragmatics: @value{GDBN} has not simply accepted the warnings
5247 enabled by @samp{-Wall -Werror -W...}. Instead it is selecting warnings
5248 when and where their benefits can be demonstrated.}
5250 @subsection Formatting
5252 @cindex source code formatting
5253 The standard GNU recommendations for formatting must be followed
5256 A function declaration should not have its name in column zero. A
5257 function definition should have its name in column zero.
5261 static void foo (void);
5269 @emph{Pragmatics: This simplifies scripting. Function definitions can
5270 be found using @samp{^function-name}.}
5272 There must be a space between a function or macro name and the opening
5273 parenthesis of its argument list (except for macro definitions, as
5274 required by C). There must not be a space after an open paren/bracket
5275 or before a close paren/bracket.
5277 While additional whitespace is generally helpful for reading, do not use
5278 more than one blank line to separate blocks, and avoid adding whitespace
5279 after the end of a program line (as of 1/99, some 600 lines had
5280 whitespace after the semicolon). Excess whitespace causes difficulties
5281 for @code{diff} and @code{patch} utilities.
5283 Pointers are declared using the traditional K&R C style:
5297 @subsection Comments
5299 @cindex comment formatting
5300 The standard GNU requirements on comments must be followed strictly.
5302 Block comments must appear in the following form, with no @code{/*}- or
5303 @code{*/}-only lines, and no leading @code{*}:
5306 /* Wait for control to return from inferior to debugger. If inferior
5307 gets a signal, we may decide to start it up again instead of
5308 returning. That is why there is a loop in this function. When
5309 this function actually returns it means the inferior should be left
5310 stopped and @value{GDBN} should read more commands. */
5313 (Note that this format is encouraged by Emacs; tabbing for a multi-line
5314 comment works correctly, and @kbd{M-q} fills the block consistently.)
5316 Put a blank line between the block comments preceding function or
5317 variable definitions, and the definition itself.
5319 In general, put function-body comments on lines by themselves, rather
5320 than trying to fit them into the 20 characters left at the end of a
5321 line, since either the comment or the code will inevitably get longer
5322 than will fit, and then somebody will have to move it anyhow.
5326 @cindex C data types
5327 Code must not depend on the sizes of C data types, the format of the
5328 host's floating point numbers, the alignment of anything, or the order
5329 of evaluation of expressions.
5331 @cindex function usage
5332 Use functions freely. There are only a handful of compute-bound areas
5333 in @value{GDBN} that might be affected by the overhead of a function
5334 call, mainly in symbol reading. Most of @value{GDBN}'s performance is
5335 limited by the target interface (whether serial line or system call).
5337 However, use functions with moderation. A thousand one-line functions
5338 are just as hard to understand as a single thousand-line function.
5340 @emph{Macros are bad, M'kay.}
5341 (But if you have to use a macro, make sure that the macro arguments are
5342 protected with parentheses.)
5346 Declarations like @samp{struct foo *} should be used in preference to
5347 declarations like @samp{typedef struct foo @{ @dots{} @} *foo_ptr}.
5350 @subsection Function Prototypes
5351 @cindex function prototypes
5353 Prototypes must be used when both @emph{declaring} and @emph{defining}
5354 a function. Prototypes for @value{GDBN} functions must include both the
5355 argument type and name, with the name matching that used in the actual
5356 function definition.
5358 All external functions should have a declaration in a header file that
5359 callers include, except for @code{_initialize_*} functions, which must
5360 be external so that @file{init.c} construction works, but shouldn't be
5361 visible to random source files.
5363 Where a source file needs a forward declaration of a static function,
5364 that declaration must appear in a block near the top of the source file.
5367 @subsection Internal Error Recovery
5369 During its execution, @value{GDBN} can encounter two types of errors.
5370 User errors and internal errors. User errors include not only a user
5371 entering an incorrect command but also problems arising from corrupt
5372 object files and system errors when interacting with the target.
5373 Internal errors include situations where @value{GDBN} has detected, at
5374 run time, a corrupt or erroneous situation.
5376 When reporting an internal error, @value{GDBN} uses
5377 @code{internal_error} and @code{gdb_assert}.
5379 @value{GDBN} must not call @code{abort} or @code{assert}.
5381 @emph{Pragmatics: There is no @code{internal_warning} function. Either
5382 the code detected a user error, recovered from it and issued a
5383 @code{warning} or the code failed to correctly recover from the user
5384 error and issued an @code{internal_error}.}
5386 @subsection File Names
5388 Any file used when building the core of @value{GDBN} must be in lower
5389 case. Any file used when building the core of @value{GDBN} must be 8.3
5390 unique. These requirements apply to both source and generated files.
5392 @emph{Pragmatics: The core of @value{GDBN} must be buildable on many
5393 platforms including DJGPP and MacOS/HFS. Every time an unfriendly file
5394 is introduced to the build process both @file{Makefile.in} and
5395 @file{configure.in} need to be modified accordingly. Compare the
5396 convoluted conversion process needed to transform @file{COPYING} into
5397 @file{copying.c} with the conversion needed to transform
5398 @file{version.in} into @file{version.c}.}
5400 Any file non 8.3 compliant file (that is not used when building the core
5401 of @value{GDBN}) must be added to @file{gdb/config/djgpp/fnchange.lst}.
5403 @emph{Pragmatics: This is clearly a compromise.}
5405 When @value{GDBN} has a local version of a system header file (ex
5406 @file{string.h}) the file name based on the POSIX header prefixed with
5407 @file{gdb_} (@file{gdb_string.h}). These headers should be relatively
5408 independent: they should use only macros defined by @file{configure},
5409 the compiler, or the host; they should include only system headers; they
5410 should refer only to system types. They may be shared between multiple
5411 programs, e.g.@: @value{GDBN} and @sc{gdbserver}.
5413 For other files @samp{-} is used as the separator.
5416 @subsection Include Files
5418 A @file{.c} file should include @file{defs.h} first.
5420 A @file{.c} file should directly include the @code{.h} file of every
5421 declaration and/or definition it directly refers to. It cannot rely on
5424 A @file{.h} file should directly include the @code{.h} file of every
5425 declaration and/or definition it directly refers to. It cannot rely on
5426 indirect inclusion. Exception: The file @file{defs.h} does not need to
5427 be directly included.
5429 An external declaration should only appear in one include file.
5431 An external declaration should never appear in a @code{.c} file.
5432 Exception: a declaration for the @code{_initialize} function that
5433 pacifies @option{-Wmissing-declaration}.
5435 A @code{typedef} definition should only appear in one include file.
5437 An opaque @code{struct} declaration can appear in multiple @file{.h}
5438 files. Where possible, a @file{.h} file should use an opaque
5439 @code{struct} declaration instead of an include.
5441 All @file{.h} files should be wrapped in:
5444 #ifndef INCLUDE_FILE_NAME_H
5445 #define INCLUDE_FILE_NAME_H
5451 @subsection Clean Design and Portable Implementation
5454 In addition to getting the syntax right, there's the little question of
5455 semantics. Some things are done in certain ways in @value{GDBN} because long
5456 experience has shown that the more obvious ways caused various kinds of
5459 @cindex assumptions about targets
5460 You can't assume the byte order of anything that comes from a target
5461 (including @var{value}s, object files, and instructions). Such things
5462 must be byte-swapped using @code{SWAP_TARGET_AND_HOST} in
5463 @value{GDBN}, or one of the swap routines defined in @file{bfd.h},
5464 such as @code{bfd_get_32}.
5466 You can't assume that you know what interface is being used to talk to
5467 the target system. All references to the target must go through the
5468 current @code{target_ops} vector.
5470 You can't assume that the host and target machines are the same machine
5471 (except in the ``native'' support modules). In particular, you can't
5472 assume that the target machine's header files will be available on the
5473 host machine. Target code must bring along its own header files --
5474 written from scratch or explicitly donated by their owner, to avoid
5478 Insertion of new @code{#ifdef}'s will be frowned upon. It's much better
5479 to write the code portably than to conditionalize it for various
5482 @cindex system dependencies
5483 New @code{#ifdef}'s which test for specific compilers or manufacturers
5484 or operating systems are unacceptable. All @code{#ifdef}'s should test
5485 for features. The information about which configurations contain which
5486 features should be segregated into the configuration files. Experience
5487 has proven far too often that a feature unique to one particular system
5488 often creeps into other systems; and that a conditional based on some
5489 predefined macro for your current system will become worthless over
5490 time, as new versions of your system come out that behave differently
5491 with regard to this feature.
5493 Adding code that handles specific architectures, operating systems,
5494 target interfaces, or hosts, is not acceptable in generic code.
5496 @cindex portable file name handling
5497 @cindex file names, portability
5498 One particularly notorious area where system dependencies tend to
5499 creep in is handling of file names. The mainline @value{GDBN} code
5500 assumes Posix semantics of file names: absolute file names begin with
5501 a forward slash @file{/}, slashes are used to separate leading
5502 directories, case-sensitive file names. These assumptions are not
5503 necessarily true on non-Posix systems such as MS-Windows. To avoid
5504 system-dependent code where you need to take apart or construct a file
5505 name, use the following portable macros:
5508 @findex HAVE_DOS_BASED_FILE_SYSTEM
5509 @item HAVE_DOS_BASED_FILE_SYSTEM
5510 This preprocessing symbol is defined to a non-zero value on hosts
5511 whose filesystems belong to the MS-DOS/MS-Windows family. Use this
5512 symbol to write conditional code which should only be compiled for
5515 @findex IS_DIR_SEPARATOR
5516 @item IS_DIR_SEPARATOR (@var{c})
5517 Evaluates to a non-zero value if @var{c} is a directory separator
5518 character. On Unix and GNU/Linux systems, only a slash @file{/} is
5519 such a character, but on Windows, both @file{/} and @file{\} will
5522 @findex IS_ABSOLUTE_PATH
5523 @item IS_ABSOLUTE_PATH (@var{file})
5524 Evaluates to a non-zero value if @var{file} is an absolute file name.
5525 For Unix and GNU/Linux hosts, a name which begins with a slash
5526 @file{/} is absolute. On DOS and Windows, @file{d:/foo} and
5527 @file{x:\bar} are also absolute file names.
5529 @findex FILENAME_CMP
5530 @item FILENAME_CMP (@var{f1}, @var{f2})
5531 Calls a function which compares file names @var{f1} and @var{f2} as
5532 appropriate for the underlying host filesystem. For Posix systems,
5533 this simply calls @code{strcmp}; on case-insensitive filesystems it
5534 will call @code{strcasecmp} instead.
5536 @findex DIRNAME_SEPARATOR
5537 @item DIRNAME_SEPARATOR
5538 Evaluates to a character which separates directories in
5539 @code{PATH}-style lists, typically held in environment variables.
5540 This character is @samp{:} on Unix, @samp{;} on DOS and Windows.
5542 @findex SLASH_STRING
5544 This evaluates to a constant string you should use to produce an
5545 absolute filename from leading directories and the file's basename.
5546 @code{SLASH_STRING} is @code{"/"} on most systems, but might be
5547 @code{"\\"} for some Windows-based ports.
5550 In addition to using these macros, be sure to use portable library
5551 functions whenever possible. For example, to extract a directory or a
5552 basename part from a file name, use the @code{dirname} and
5553 @code{basename} library functions (available in @code{libiberty} for
5554 platforms which don't provide them), instead of searching for a slash
5555 with @code{strrchr}.
5557 Another way to generalize @value{GDBN} along a particular interface is with an
5558 attribute struct. For example, @value{GDBN} has been generalized to handle
5559 multiple kinds of remote interfaces---not by @code{#ifdef}s everywhere, but
5560 by defining the @code{target_ops} structure and having a current target (as
5561 well as a stack of targets below it, for memory references). Whenever
5562 something needs to be done that depends on which remote interface we are
5563 using, a flag in the current target_ops structure is tested (e.g.,
5564 @code{target_has_stack}), or a function is called through a pointer in the
5565 current target_ops structure. In this way, when a new remote interface
5566 is added, only one module needs to be touched---the one that actually
5567 implements the new remote interface. Other examples of
5568 attribute-structs are BFD access to multiple kinds of object file
5569 formats, or @value{GDBN}'s access to multiple source languages.
5571 Please avoid duplicating code. For example, in @value{GDBN} 3.x all
5572 the code interfacing between @code{ptrace} and the rest of
5573 @value{GDBN} was duplicated in @file{*-dep.c}, and so changing
5574 something was very painful. In @value{GDBN} 4.x, these have all been
5575 consolidated into @file{infptrace.c}. @file{infptrace.c} can deal
5576 with variations between systems the same way any system-independent
5577 file would (hooks, @code{#if defined}, etc.), and machines which are
5578 radically different don't need to use @file{infptrace.c} at all.
5580 All debugging code must be controllable using the @samp{set debug
5581 @var{module}} command. Do not use @code{printf} to print trace
5582 messages. Use @code{fprintf_unfiltered(gdb_stdlog, ...}. Do not use
5583 @code{#ifdef DEBUG}.
5588 @chapter Porting @value{GDBN}
5589 @cindex porting to new machines
5591 Most of the work in making @value{GDBN} compile on a new machine is in
5592 specifying the configuration of the machine. This is done in a
5593 dizzying variety of header files and configuration scripts, which we
5594 hope to make more sensible soon. Let's say your new host is called an
5595 @var{xyz} (e.g., @samp{sun4}), and its full three-part configuration
5596 name is @code{@var{arch}-@var{xvend}-@var{xos}} (e.g.,
5597 @samp{sparc-sun-sunos4}). In particular:
5601 In the top level directory, edit @file{config.sub} and add @var{arch},
5602 @var{xvend}, and @var{xos} to the lists of supported architectures,
5603 vendors, and operating systems near the bottom of the file. Also, add
5604 @var{xyz} as an alias that maps to
5605 @code{@var{arch}-@var{xvend}-@var{xos}}. You can test your changes by
5609 ./config.sub @var{xyz}
5616 ./config.sub @code{@var{arch}-@var{xvend}-@var{xos}}
5620 which should both respond with @code{@var{arch}-@var{xvend}-@var{xos}}
5621 and no error messages.
5624 You need to port BFD, if that hasn't been done already. Porting BFD is
5625 beyond the scope of this manual.
5628 To configure @value{GDBN} itself, edit @file{gdb/configure.host} to recognize
5629 your system and set @code{gdb_host} to @var{xyz}, and (unless your
5630 desired target is already available) also edit @file{gdb/configure.tgt},
5631 setting @code{gdb_target} to something appropriate (for instance,
5634 @emph{Maintainer's note: Work in progress. The file
5635 @file{gdb/configure.host} originally needed to be modified when either a
5636 new native target or a new host machine was being added to @value{GDBN}.
5637 Recent changes have removed this requirement. The file now only needs
5638 to be modified when adding a new native configuration. This will likely
5639 changed again in the future.}
5642 Finally, you'll need to specify and define @value{GDBN}'s host-, native-, and
5643 target-dependent @file{.h} and @file{.c} files used for your
5647 @node Versions and Branches
5648 @chapter Versions and Branches
5652 @value{GDBN}'s version is determined by the file
5653 @file{gdb/version.in} and takes one of the following forms:
5656 @item @var{major}.@var{minor}
5657 @itemx @var{major}.@var{minor}.@var{patchlevel}
5658 an official release (e.g., 6.2 or 6.2.1)
5659 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}
5660 a snapshot taken at @var{YYYY}-@var{MM}-@var{DD}-gmt (e.g.,
5661 6.1.50.20020302, 6.1.90.20020304, or 6.1.0.20020308)
5662 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD}-cvs
5663 a @sc{cvs} check out drawn on @var{YYYY}-@var{MM}-@var{DD} (e.g.,
5664 6.1.50.20020302-cvs, 6.1.90.20020304-cvs, or 6.1.0.20020308-cvs)
5665 @item @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD} (@var{vendor})
5666 a vendor specific release of @value{GDBN}, that while based on@*
5667 @var{major}.@var{minor}.@var{patchlevel}.@var{YYYY}@var{MM}@var{DD},
5668 may include additional changes
5671 @value{GDBN}'s mainline uses the @var{major} and @var{minor} version
5672 numbers from the most recent release branch, with a @var{patchlevel}
5673 of 50. At the time each new release branch is created, the mainline's
5674 @var{major} and @var{minor} version numbers are updated.
5676 @value{GDBN}'s release branch is similar. When the branch is cut, the
5677 @var{patchlevel} is changed from 50 to 90. As draft releases are
5678 drawn from the branch, the @var{patchlevel} is incremented. Once the
5679 first release (@var{major}.@var{minor}) has been made, the
5680 @var{patchlevel} is set to 0 and updates have an incremented
5683 For snapshots, and @sc{cvs} check outs, it is also possible to
5684 identify the @sc{cvs} origin:
5687 @item @var{major}.@var{minor}.50.@var{YYYY}@var{MM}@var{DD}
5688 drawn from the @sc{head} of mainline @sc{cvs} (e.g., 6.1.50.20020302)
5689 @item @var{major}.@var{minor}.90.@var{YYYY}@var{MM}@var{DD}
5690 @itemx @var{major}.@var{minor}.91.@var{YYYY}@var{MM}@var{DD} @dots{}
5691 drawn from a release branch prior to the release (e.g.,
5693 @item @var{major}.@var{minor}.0.@var{YYYY}@var{MM}@var{DD}
5694 @itemx @var{major}.@var{minor}.1.@var{YYYY}@var{MM}@var{DD} @dots{}
5695 drawn from a release branch after the release (e.g., 6.2.0.20020308)
5698 If the previous @value{GDBN} version is 6.1 and the current version is
5699 6.2, then, substituting 6 for @var{major} and 1 or 2 for @var{minor},
5700 here's an illustration of a typical sequence:
5707 +--------------------------.
5710 6.2.50.20020303-cvs 6.1.90 (draft #1)
5712 6.2.50.20020304-cvs 6.1.90.20020304-cvs
5714 6.2.50.20020305-cvs 6.1.91 (draft #2)
5716 6.2.50.20020306-cvs 6.1.91.20020306-cvs
5718 6.2.50.20020307-cvs 6.2 (release)
5720 6.2.50.20020308-cvs 6.2.0.20020308-cvs
5722 6.2.50.20020309-cvs 6.2.1 (update)
5724 6.2.50.20020310-cvs <branch closed>
5728 +--------------------------.
5731 6.3.50.20020312-cvs 6.2.90 (draft #1)
5735 @section Release Branches
5736 @cindex Release Branches
5738 @value{GDBN} draws a release series (6.2, 6.2.1, @dots{}) from a
5739 single release branch, and identifies that branch using the @sc{cvs}
5743 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-branchpoint
5744 gdb_@var{major}_@var{minor}-branch
5745 gdb_@var{major}_@var{minor}-@var{YYYY}@var{MM}@var{DD}-release
5748 @emph{Pragmatics: To help identify the date at which a branch or
5749 release is made, both the branchpoint and release tags include the
5750 date that they are cut (@var{YYYY}@var{MM}@var{DD}) in the tag. The
5751 branch tag, denoting the head of the branch, does not need this.}
5753 @section Vendor Branches
5754 @cindex vendor branches
5756 To avoid version conflicts, vendors are expected to modify the file
5757 @file{gdb/version.in} to include a vendor unique alphabetic identifier
5758 (an official @value{GDBN} release never uses alphabetic characters in
5759 its version identifer). E.g., @samp{6.2widgit2}, or @samp{6.2 (Widgit
5762 @section Experimental Branches
5763 @cindex experimental branches
5765 @subsection Guidelines
5767 @value{GDBN} permits the creation of branches, cut from the @sc{cvs}
5768 repository, for experimental development. Branches make it possible
5769 for developers to share preliminary work, and maintainers to examine
5770 significant new developments.
5772 The following are a set of guidelines for creating such branches:
5776 @item a branch has an owner
5777 The owner can set further policy for a branch, but may not change the
5778 ground rules. In particular, they can set a policy for commits (be it
5779 adding more reviewers or deciding who can commit).
5781 @item all commits are posted
5782 All changes committed to a branch shall also be posted to
5783 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} patches
5784 mailing list}. While commentary on such changes are encouraged, people
5785 should remember that the changes only apply to a branch.
5787 @item all commits are covered by an assignment
5788 This ensures that all changes belong to the Free Software Foundation,
5789 and avoids the possibility that the branch may become contaminated.
5791 @item a branch is focused
5792 A focused branch has a single objective or goal, and does not contain
5793 unnecessary or irrelevant changes. Cleanups, where identified, being
5794 be pushed into the mainline as soon as possible.
5796 @item a branch tracks mainline
5797 This keeps the level of divergence under control. It also keeps the
5798 pressure on developers to push cleanups and other stuff into the
5801 @item a branch shall contain the entire @value{GDBN} module
5802 The @value{GDBN} module @code{gdb} should be specified when creating a
5803 branch (branches of individual files should be avoided). @xref{Tags}.
5805 @item a branch shall be branded using @file{version.in}
5806 The file @file{gdb/version.in} shall be modified so that it identifies
5807 the branch @var{owner} and branch @var{name}, e.g.,
5808 @samp{6.2.50.20030303_owner_name} or @samp{6.2 (Owner Name)}.
5815 To simplify the identification of @value{GDBN} branches, the following
5816 branch tagging convention is strongly recommended:
5820 @item @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5821 @itemx @var{owner}_@var{name}-@var{YYYYMMDD}-branch
5822 The branch point and corresponding branch tag. @var{YYYYMMDD} is the
5823 date that the branch was created. A branch is created using the
5824 sequence: @anchor{experimental branch tags}
5826 cvs rtag @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint gdb
5827 cvs rtag -b -r @var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint \
5828 @var{owner}_@var{name}-@var{YYYYMMDD}-branch gdb
5831 @item @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5832 The tagged point, on the mainline, that was used when merging the branch
5833 on @var{yyyymmdd}. To merge in all changes since the branch was cut,
5834 use a command sequence like:
5836 cvs rtag @var{owner}_@var{name}-@var{yyyymmdd}-mergepoint gdb
5838 -j@var{owner}_@var{name}-@var{YYYYMMDD}-branchpoint
5839 -j@var{owner}_@var{name}-@var{yyyymmdd}-mergepoint
5842 Similar sequences can be used to just merge in changes since the last
5848 For further information on @sc{cvs}, see
5849 @uref{http://www.gnu.org/software/cvs/, Concurrent Versions System}.
5851 @node Start of New Year Procedure
5852 @chapter Start of New Year Procedure
5853 @cindex new year procedure
5855 At the start of each new year, the following actions should be performed:
5859 Rotate the ChangeLog file
5861 The current @file{ChangeLog} file should be renamed into
5862 @file{ChangeLog-YYYY} where YYYY is the year that has just passed.
5863 A new @file{ChangeLog} file should be created, and its contents should
5864 contain a reference to the previous ChangeLog. The following should
5865 also be preserved at the end of the new ChangeLog, in order to provide
5866 the appropriate settings when editing this file with Emacs:
5872 version-control: never
5877 Update the copyright year in the startup message
5879 Update the copyright year in file @file{top.c}, function
5880 @code{print_gdb_version}.
5885 @chapter Releasing @value{GDBN}
5886 @cindex making a new release of gdb
5888 @section Branch Commit Policy
5890 The branch commit policy is pretty slack. @value{GDBN} releases 5.0,
5891 5.1 and 5.2 all used the below:
5895 The @file{gdb/MAINTAINERS} file still holds.
5897 Don't fix something on the branch unless/until it is also fixed in the
5898 trunk. If this isn't possible, mentioning it in the @file{gdb/PROBLEMS}
5899 file is better than committing a hack.
5901 When considering a patch for the branch, suggested criteria include:
5902 Does it fix a build? Does it fix the sequence @kbd{break main; run}
5903 when debugging a static binary?
5905 The further a change is from the core of @value{GDBN}, the less likely
5906 the change will worry anyone (e.g., target specific code).
5908 Only post a proposal to change the core of @value{GDBN} after you've
5909 sent individual bribes to all the people listed in the
5910 @file{MAINTAINERS} file @t{;-)}
5913 @emph{Pragmatics: Provided updates are restricted to non-core
5914 functionality there is little chance that a broken change will be fatal.
5915 This means that changes such as adding a new architectures or (within
5916 reason) support for a new host are considered acceptable.}
5919 @section Obsoleting code
5921 Before anything else, poke the other developers (and around the source
5922 code) to see if there is anything that can be removed from @value{GDBN}
5923 (an old target, an unused file).
5925 Obsolete code is identified by adding an @code{OBSOLETE} prefix to every
5926 line. Doing this means that it is easy to identify something that has
5927 been obsoleted when greping through the sources.
5929 The process is done in stages --- this is mainly to ensure that the
5930 wider @value{GDBN} community has a reasonable opportunity to respond.
5931 Remember, everything on the Internet takes a week.
5935 Post the proposal on @email{gdb@@sources.redhat.com, the GDB mailing
5936 list} Creating a bug report to track the task's state, is also highly
5941 Post the proposal on @email{gdb-announce@@sources.redhat.com, the GDB
5942 Announcement mailing list}.
5946 Go through and edit all relevant files and lines so that they are
5947 prefixed with the word @code{OBSOLETE}.
5949 Wait until the next GDB version, containing this obsolete code, has been
5952 Remove the obsolete code.
5956 @emph{Maintainer note: While removing old code is regrettable it is
5957 hopefully better for @value{GDBN}'s long term development. Firstly it
5958 helps the developers by removing code that is either no longer relevant
5959 or simply wrong. Secondly since it removes any history associated with
5960 the file (effectively clearing the slate) the developer has a much freer
5961 hand when it comes to fixing broken files.}
5965 @section Before the Branch
5967 The most important objective at this stage is to find and fix simple
5968 changes that become a pain to track once the branch is created. For
5969 instance, configuration problems that stop @value{GDBN} from even
5970 building. If you can't get the problem fixed, document it in the
5971 @file{gdb/PROBLEMS} file.
5973 @subheading Prompt for @file{gdb/NEWS}
5975 People always forget. Send a post reminding them but also if you know
5976 something interesting happened add it yourself. The @code{schedule}
5977 script will mention this in its e-mail.
5979 @subheading Review @file{gdb/README}
5981 Grab one of the nightly snapshots and then walk through the
5982 @file{gdb/README} looking for anything that can be improved. The
5983 @code{schedule} script will mention this in its e-mail.
5985 @subheading Refresh any imported files.
5987 A number of files are taken from external repositories. They include:
5991 @file{texinfo/texinfo.tex}
5993 @file{config.guess} et.@: al.@: (see the top-level @file{MAINTAINERS}
5996 @file{etc/standards.texi}, @file{etc/make-stds.texi}
5999 @subheading Check the ARI
6001 @uref{http://sources.redhat.com/gdb/ari,,A.R.I.} is an @code{awk} script
6002 (Awk Regression Index ;-) that checks for a number of errors and coding
6003 conventions. The checks include things like using @code{malloc} instead
6004 of @code{xmalloc} and file naming problems. There shouldn't be any
6007 @subsection Review the bug data base
6009 Close anything obviously fixed.
6011 @subsection Check all cross targets build
6013 The targets are listed in @file{gdb/MAINTAINERS}.
6016 @section Cut the Branch
6018 @subheading Create the branch
6023 $ V=`echo $v | sed 's/\./_/g'`
6024 $ D=`date -u +%Y-%m-%d`
6027 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6028 -D $D-gmt gdb_$V-$D-branchpoint insight+dejagnu
6029 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag
6030 -D 2002-03-03-gmt gdb_5_2-2002-03-03-branchpoint insight+dejagnu
6033 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6034 -b -r gdb_$V-$D-branchpoint gdb_$V-branch insight+dejagnu
6035 cvs -f -d :ext:sources.redhat.com:/cvs/src rtag \
6036 -b -r gdb_5_2-2002-03-03-branchpoint gdb_5_2-branch insight+dejagnu
6044 by using @kbd{-D YYYY-MM-DD-gmt} the branch is forced to an exact
6047 the trunk is first taged so that the branch point can easily be found
6049 Insight (which includes GDB) and dejagnu are all tagged at the same time
6051 @file{version.in} gets bumped to avoid version number conflicts
6053 the reading of @file{.cvsrc} is disabled using @file{-f}
6056 @subheading Update @file{version.in}
6061 $ V=`echo $v | sed 's/\./_/g'`
6065 $ echo cvs -f -d :ext:sources.redhat.com:/cvs/src co \
6066 -r gdb_$V-branch src/gdb/version.in
6067 cvs -f -d :ext:sources.redhat.com:/cvs/src co
6068 -r gdb_5_2-branch src/gdb/version.in
6070 U src/gdb/version.in
6072 $ echo $u.90-0000-00-00-cvs > version.in
6074 5.1.90-0000-00-00-cvs
6075 $ cvs -f commit version.in
6080 @file{0000-00-00} is used as a date to pump prime the version.in update
6083 @file{.90} and the previous branch version are used as fairly arbitrary
6084 initial branch version number
6088 @subheading Update the web and news pages
6092 @subheading Tweak cron to track the new branch
6094 The file @file{gdbadmin/cron/crontab} contains gdbadmin's cron table.
6095 This file needs to be updated so that:
6099 a daily timestamp is added to the file @file{version.in}
6101 the new branch is included in the snapshot process
6105 See the file @file{gdbadmin/cron/README} for how to install the updated
6108 The file @file{gdbadmin/ss/README} should also be reviewed to reflect
6109 any changes. That file is copied to both the branch/ and current/
6110 snapshot directories.
6113 @subheading Update the NEWS and README files
6115 The @file{NEWS} file needs to be updated so that on the branch it refers
6116 to @emph{changes in the current release} while on the trunk it also
6117 refers to @emph{changes since the current release}.
6119 The @file{README} file needs to be updated so that it refers to the
6122 @subheading Post the branch info
6124 Send an announcement to the mailing lists:
6128 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6130 @email{gdb@@sources.redhat.com, GDB Discsussion mailing list} and
6131 @email{gdb-testers@@sources.redhat.com, GDB Discsussion mailing list}
6134 @emph{Pragmatics: The branch creation is sent to the announce list to
6135 ensure that people people not subscribed to the higher volume discussion
6138 The announcement should include:
6144 how to check out the branch using CVS
6146 the date/number of weeks until the release
6148 the branch commit policy
6152 @section Stabilize the branch
6154 Something goes here.
6156 @section Create a Release
6158 The process of creating and then making available a release is broken
6159 down into a number of stages. The first part addresses the technical
6160 process of creating a releasable tar ball. The later stages address the
6161 process of releasing that tar ball.
6163 When making a release candidate just the first section is needed.
6165 @subsection Create a release candidate
6167 The objective at this stage is to create a set of tar balls that can be
6168 made available as a formal release (or as a less formal release
6171 @subsubheading Freeze the branch
6173 Send out an e-mail notifying everyone that the branch is frozen to
6174 @email{gdb-patches@@sources.redhat.com}.
6176 @subsubheading Establish a few defaults.
6181 $ t=/sourceware/snapshot-tmp/gdbadmin-tmp
6183 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6187 /sourceware/snapshot-tmp/gdbadmin-tmp/gdb_5_2-branch/5.2
6189 /home/gdbadmin/bin/autoconf
6198 Check the @code{autoconf} version carefully. You want to be using the
6199 version taken from the @file{binutils} snapshot directory, which can be
6200 found at @uref{ftp://sources.redhat.com/pub/binutils/}. It is very
6201 unlikely that a system installed version of @code{autoconf} (e.g.,
6202 @file{/usr/bin/autoconf}) is correct.
6205 @subsubheading Check out the relevant modules:
6208 $ for m in gdb insight dejagnu
6210 ( mkdir -p $m && cd $m && cvs -q -f -d /cvs/src co -P -r $b $m )
6220 The reading of @file{.cvsrc} is disabled (@file{-f}) so that there isn't
6221 any confusion between what is written here and what your local
6222 @code{cvs} really does.
6225 @subsubheading Update relevant files.
6231 Major releases get their comments added as part of the mainline. Minor
6232 releases should probably mention any significant bugs that were fixed.
6234 Don't forget to include the @file{ChangeLog} entry.
6237 $ emacs gdb/src/gdb/NEWS
6242 $ cp gdb/src/gdb/NEWS insight/src/gdb/NEWS
6243 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6248 You'll need to update:
6260 $ emacs gdb/src/gdb/README
6265 $ cp gdb/src/gdb/README insight/src/gdb/README
6266 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6269 @emph{Maintainer note: Hopefully the @file{README} file was reviewed
6270 before the initial branch was cut so just a simple substitute is needed
6273 @emph{Maintainer note: Other projects generate @file{README} and
6274 @file{INSTALL} from the core documentation. This might be worth
6277 @item gdb/version.in
6280 $ echo $v > gdb/src/gdb/version.in
6281 $ cat gdb/src/gdb/version.in
6283 $ emacs gdb/src/gdb/version.in
6286 ... Bump to version ...
6288 $ cp gdb/src/gdb/version.in insight/src/gdb/version.in
6289 $ cp gdb/src/gdb/ChangeLog insight/src/gdb/ChangeLog
6292 @item dejagnu/src/dejagnu/configure.in
6294 Dejagnu is more complicated. The version number is a parameter to
6295 @code{AM_INIT_AUTOMAKE}. Tweak it to read something like gdb-5.1.91.
6297 Don't forget to re-generate @file{configure}.
6299 Don't forget to include a @file{ChangeLog} entry.
6302 $ emacs dejagnu/src/dejagnu/configure.in
6307 $ ( cd dejagnu/src/dejagnu && autoconf )
6312 @subsubheading Do the dirty work
6314 This is identical to the process used to create the daily snapshot.
6317 $ for m in gdb insight
6319 ( cd $m/src && gmake -f src-release $m.tar )
6321 $ ( m=dejagnu; cd $m/src && gmake -f src-release $m.tar.bz2 )
6324 If the top level source directory does not have @file{src-release}
6325 (@value{GDBN} version 5.3.1 or earlier), try these commands instead:
6328 $ for m in gdb insight
6330 ( cd $m/src && gmake -f Makefile.in $m.tar )
6332 $ ( m=dejagnu; cd $m/src && gmake -f Makefile.in $m.tar.bz2 )
6335 @subsubheading Check the source files
6337 You're looking for files that have mysteriously disappeared.
6338 @kbd{distclean} has the habit of deleting files it shouldn't. Watch out
6339 for the @file{version.in} update @kbd{cronjob}.
6342 $ ( cd gdb/src && cvs -f -q -n update )
6346 @dots{} lots of generated files @dots{}
6351 @dots{} lots of generated files @dots{}
6356 @emph{Don't worry about the @file{gdb.info-??} or
6357 @file{gdb/p-exp.tab.c}. They were generated (and yes @file{gdb.info-1}
6358 was also generated only something strange with CVS means that they
6359 didn't get supressed). Fixing it would be nice though.}
6361 @subsubheading Create compressed versions of the release
6367 dejagnu/ dejagnu-gdb-5.2.tar.bz2 gdb/ gdb-5.2.tar insight/ insight-5.2.tar
6368 $ for m in gdb insight
6370 bzip2 -v -9 -c $m-$v.tar > $m-$v.tar.bz2
6371 gzip -v -9 -c $m-$v.tar > $m-$v.tar.gz
6381 A pipe such as @kbd{bunzip2 < xxx.bz2 | gzip -9 > xxx.gz} is not since,
6382 in that mode, @code{gzip} does not know the name of the file and, hence,
6383 can not include it in the compressed file. This is also why the release
6384 process runs @code{tar} and @code{bzip2} as separate passes.
6387 @subsection Sanity check the tar ball
6389 Pick a popular machine (Solaris/PPC?) and try the build on that.
6392 $ bunzip2 < gdb-5.2.tar.bz2 | tar xpf -
6397 $ ./gdb/gdb ./gdb/gdb
6401 Breakpoint 1 at 0x80732bc: file main.c, line 734.
6403 Starting program: /tmp/gdb-5.2/gdb/gdb
6405 Breakpoint 1, main (argc=1, argv=0xbffff8b4) at main.c:734
6406 734 catch_errors (captured_main, &args, "", RETURN_MASK_ALL);
6408 $1 = @{argc = 136426532, argv = 0x821b7f0@}
6412 @subsection Make a release candidate available
6414 If this is a release candidate then the only remaining steps are:
6418 Commit @file{version.in} and @file{ChangeLog}
6420 Tweak @file{version.in} (and @file{ChangeLog} to read
6421 @var{L}.@var{M}.@var{N}-0000-00-00-cvs so that the version update
6422 process can restart.
6424 Make the release candidate available in
6425 @uref{ftp://sources.redhat.com/pub/gdb/snapshots/branch}
6427 Notify the relevant mailing lists ( @email{gdb@@sources.redhat.com} and
6428 @email{gdb-testers@@sources.redhat.com} that the candidate is available.
6431 @subsection Make a formal release available
6433 (And you thought all that was required was to post an e-mail.)
6435 @subsubheading Install on sware
6437 Copy the new files to both the release and the old release directory:
6440 $ cp *.bz2 *.gz ~ftp/pub/gdb/old-releases/
6441 $ cp *.bz2 *.gz ~ftp/pub/gdb/releases
6445 Clean up the releases directory so that only the most recent releases
6446 are available (e.g. keep 5.2 and 5.2.1 but remove 5.1):
6449 $ cd ~ftp/pub/gdb/releases
6454 Update the file @file{README} and @file{.message} in the releases
6461 $ ln README .message
6464 @subsubheading Update the web pages.
6468 @item htdocs/download/ANNOUNCEMENT
6469 This file, which is posted as the official announcement, includes:
6472 General announcement
6474 News. If making an @var{M}.@var{N}.1 release, retain the news from
6475 earlier @var{M}.@var{N} release.
6480 @item htdocs/index.html
6481 @itemx htdocs/news/index.html
6482 @itemx htdocs/download/index.html
6483 These files include:
6486 announcement of the most recent release
6488 news entry (remember to update both the top level and the news directory).
6490 These pages also need to be regenerate using @code{index.sh}.
6492 @item download/onlinedocs/
6493 You need to find the magic command that is used to generate the online
6494 docs from the @file{.tar.bz2}. The best way is to look in the output
6495 from one of the nightly @code{cron} jobs and then just edit accordingly.
6499 $ ~/ss/update-web-docs \
6500 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6502 /www/sourceware/htdocs/gdb/download/onlinedocs \
6507 Just like the online documentation. Something like:
6510 $ /bin/sh ~/ss/update-web-ari \
6511 ~ftp/pub/gdb/releases/gdb-5.2.tar.bz2 \
6513 /www/sourceware/htdocs/gdb/download/ari \
6519 @subsubheading Shadow the pages onto gnu
6521 Something goes here.
6524 @subsubheading Install the @value{GDBN} tar ball on GNU
6526 At the time of writing, the GNU machine was @kbd{gnudist.gnu.org} in
6527 @file{~ftp/gnu/gdb}.
6529 @subsubheading Make the @file{ANNOUNCEMENT}
6531 Post the @file{ANNOUNCEMENT} file you created above to:
6535 @email{gdb-announce@@sources.redhat.com, GDB Announcement mailing list}
6537 @email{info-gnu@@gnu.org, General GNU Announcement list} (but delay it a
6538 day or so to let things get out)
6540 @email{bug-gdb@@gnu.org, GDB Bug Report mailing list}
6545 The release is out but you're still not finished.
6547 @subsubheading Commit outstanding changes
6549 In particular you'll need to commit any changes to:
6553 @file{gdb/ChangeLog}
6555 @file{gdb/version.in}
6562 @subsubheading Tag the release
6567 $ d=`date -u +%Y-%m-%d`
6570 $ ( cd insight/src/gdb && cvs -f -q update )
6571 $ ( cd insight/src && cvs -f -q tag gdb_5_2-$d-release )
6574 Insight is used since that contains more of the release than
6575 @value{GDBN} (@code{dejagnu} doesn't get tagged but I think we can live
6578 @subsubheading Mention the release on the trunk
6580 Just put something in the @file{ChangeLog} so that the trunk also
6581 indicates when the release was made.
6583 @subsubheading Restart @file{gdb/version.in}
6585 If @file{gdb/version.in} does not contain an ISO date such as
6586 @kbd{2002-01-24} then the daily @code{cronjob} won't update it. Having
6587 committed all the release changes it can be set to
6588 @file{5.2.0_0000-00-00-cvs} which will restart things (yes the @kbd{_}
6589 is important - it affects the snapshot process).
6591 Don't forget the @file{ChangeLog}.
6593 @subsubheading Merge into trunk
6595 The files committed to the branch may also need changes merged into the
6598 @subsubheading Revise the release schedule
6600 Post a revised release schedule to @email{gdb@@sources.redhat.com, GDB
6601 Discussion List} with an updated announcement. The schedule can be
6602 generated by running:
6605 $ ~/ss/schedule `date +%s` schedule
6609 The first parameter is approximate date/time in seconds (from the epoch)
6610 of the most recent release.
6612 Also update the schedule @code{cronjob}.
6614 @section Post release
6616 Remove any @code{OBSOLETE} code.
6623 The testsuite is an important component of the @value{GDBN} package.
6624 While it is always worthwhile to encourage user testing, in practice
6625 this is rarely sufficient; users typically use only a small subset of
6626 the available commands, and it has proven all too common for a change
6627 to cause a significant regression that went unnoticed for some time.
6629 The @value{GDBN} testsuite uses the DejaGNU testing framework.
6630 DejaGNU is built using @code{Tcl} and @code{expect}. The tests
6631 themselves are calls to various @code{Tcl} procs; the framework runs all the
6632 procs and summarizes the passes and fails.
6634 @section Using the Testsuite
6636 @cindex running the test suite
6637 To run the testsuite, simply go to the @value{GDBN} object directory (or to the
6638 testsuite's objdir) and type @code{make check}. This just sets up some
6639 environment variables and invokes DejaGNU's @code{runtest} script. While
6640 the testsuite is running, you'll get mentions of which test file is in use,
6641 and a mention of any unexpected passes or fails. When the testsuite is
6642 finished, you'll get a summary that looks like this:
6647 # of expected passes 6016
6648 # of unexpected failures 58
6649 # of unexpected successes 5
6650 # of expected failures 183
6651 # of unresolved testcases 3
6652 # of untested testcases 5
6655 To run a specific test script, type:
6657 make check RUNTESTFLAGS='@var{tests}'
6659 where @var{tests} is a list of test script file names, separated by
6662 The ideal test run consists of expected passes only; however, reality
6663 conspires to keep us from this ideal. Unexpected failures indicate
6664 real problems, whether in @value{GDBN} or in the testsuite. Expected
6665 failures are still failures, but ones which have been decided are too
6666 hard to deal with at the time; for instance, a test case might work
6667 everywhere except on AIX, and there is no prospect of the AIX case
6668 being fixed in the near future. Expected failures should not be added
6669 lightly, since you may be masking serious bugs in @value{GDBN}.
6670 Unexpected successes are expected fails that are passing for some
6671 reason, while unresolved and untested cases often indicate some minor
6672 catastrophe, such as the compiler being unable to deal with a test
6675 When making any significant change to @value{GDBN}, you should run the
6676 testsuite before and after the change, to confirm that there are no
6677 regressions. Note that truly complete testing would require that you
6678 run the testsuite with all supported configurations and a variety of
6679 compilers; however this is more than really necessary. In many cases
6680 testing with a single configuration is sufficient. Other useful
6681 options are to test one big-endian (Sparc) and one little-endian (x86)
6682 host, a cross config with a builtin simulator (powerpc-eabi,
6683 mips-elf), or a 64-bit host (Alpha).
6685 If you add new functionality to @value{GDBN}, please consider adding
6686 tests for it as well; this way future @value{GDBN} hackers can detect
6687 and fix their changes that break the functionality you added.
6688 Similarly, if you fix a bug that was not previously reported as a test
6689 failure, please add a test case for it. Some cases are extremely
6690 difficult to test, such as code that handles host OS failures or bugs
6691 in particular versions of compilers, and it's OK not to try to write
6692 tests for all of those.
6694 DejaGNU supports separate build, host, and target machines. However,
6695 some @value{GDBN} test scripts do not work if the build machine and
6696 the host machine are not the same. In such an environment, these scripts
6697 will give a result of ``UNRESOLVED'', like this:
6700 UNRESOLVED: gdb.base/example.exp: This test script does not work on a remote host.
6703 @section Testsuite Organization
6705 @cindex test suite organization
6706 The testsuite is entirely contained in @file{gdb/testsuite}. While the
6707 testsuite includes some makefiles and configury, these are very minimal,
6708 and used for little besides cleaning up, since the tests themselves
6709 handle the compilation of the programs that @value{GDBN} will run. The file
6710 @file{testsuite/lib/gdb.exp} contains common utility procs useful for
6711 all @value{GDBN} tests, while the directory @file{testsuite/config} contains
6712 configuration-specific files, typically used for special-purpose
6713 definitions of procs like @code{gdb_load} and @code{gdb_start}.
6715 The tests themselves are to be found in @file{testsuite/gdb.*} and
6716 subdirectories of those. The names of the test files must always end
6717 with @file{.exp}. DejaGNU collects the test files by wildcarding
6718 in the test directories, so both subdirectories and individual files
6719 get chosen and run in alphabetical order.
6721 The following table lists the main types of subdirectories and what they
6722 are for. Since DejaGNU finds test files no matter where they are
6723 located, and since each test file sets up its own compilation and
6724 execution environment, this organization is simply for convenience and
6729 This is the base testsuite. The tests in it should apply to all
6730 configurations of @value{GDBN} (but generic native-only tests may live here).
6731 The test programs should be in the subset of C that is valid K&R,
6732 ANSI/ISO, and C@t{++} (@code{#ifdef}s are allowed if necessary, for instance
6735 @item gdb.@var{lang}
6736 Language-specific tests for any language @var{lang} besides C. Examples are
6737 @file{gdb.cp} and @file{gdb.java}.
6739 @item gdb.@var{platform}
6740 Non-portable tests. The tests are specific to a specific configuration
6741 (host or target), such as HP-UX or eCos. Example is @file{gdb.hp}, for
6744 @item gdb.@var{compiler}
6745 Tests specific to a particular compiler. As of this writing (June
6746 1999), there aren't currently any groups of tests in this category that
6747 couldn't just as sensibly be made platform-specific, but one could
6748 imagine a @file{gdb.gcc}, for tests of @value{GDBN}'s handling of GCC
6751 @item gdb.@var{subsystem}
6752 Tests that exercise a specific @value{GDBN} subsystem in more depth. For
6753 instance, @file{gdb.disasm} exercises various disassemblers, while
6754 @file{gdb.stabs} tests pathways through the stabs symbol reader.
6757 @section Writing Tests
6758 @cindex writing tests
6760 In many areas, the @value{GDBN} tests are already quite comprehensive; you
6761 should be able to copy existing tests to handle new cases.
6763 You should try to use @code{gdb_test} whenever possible, since it
6764 includes cases to handle all the unexpected errors that might happen.
6765 However, it doesn't cost anything to add new test procedures; for
6766 instance, @file{gdb.base/exprs.exp} defines a @code{test_expr} that
6767 calls @code{gdb_test} multiple times.
6769 Only use @code{send_gdb} and @code{gdb_expect} when absolutely
6770 necessary, such as when @value{GDBN} has several valid responses to a command.
6772 The source language programs do @emph{not} need to be in a consistent
6773 style. Since @value{GDBN} is used to debug programs written in many different
6774 styles, it's worth having a mix of styles in the testsuite; for
6775 instance, some @value{GDBN} bugs involving the display of source lines would
6776 never manifest themselves if the programs used GNU coding style
6783 Check the @file{README} file, it often has useful information that does not
6784 appear anywhere else in the directory.
6787 * Getting Started:: Getting started working on @value{GDBN}
6788 * Debugging GDB:: Debugging @value{GDBN} with itself
6791 @node Getting Started,,, Hints
6793 @section Getting Started
6795 @value{GDBN} is a large and complicated program, and if you first starting to
6796 work on it, it can be hard to know where to start. Fortunately, if you
6797 know how to go about it, there are ways to figure out what is going on.
6799 This manual, the @value{GDBN} Internals manual, has information which applies
6800 generally to many parts of @value{GDBN}.
6802 Information about particular functions or data structures are located in
6803 comments with those functions or data structures. If you run across a
6804 function or a global variable which does not have a comment correctly
6805 explaining what is does, this can be thought of as a bug in @value{GDBN}; feel
6806 free to submit a bug report, with a suggested comment if you can figure
6807 out what the comment should say. If you find a comment which is
6808 actually wrong, be especially sure to report that.
6810 Comments explaining the function of macros defined in host, target, or
6811 native dependent files can be in several places. Sometimes they are
6812 repeated every place the macro is defined. Sometimes they are where the
6813 macro is used. Sometimes there is a header file which supplies a
6814 default definition of the macro, and the comment is there. This manual
6815 also documents all the available macros.
6816 @c (@pxref{Host Conditionals}, @pxref{Target
6817 @c Conditionals}, @pxref{Native Conditionals}, and @pxref{Obsolete
6820 Start with the header files. Once you have some idea of how
6821 @value{GDBN}'s internal symbol tables are stored (see @file{symtab.h},
6822 @file{gdbtypes.h}), you will find it much easier to understand the
6823 code which uses and creates those symbol tables.
6825 You may wish to process the information you are getting somehow, to
6826 enhance your understanding of it. Summarize it, translate it to another
6827 language, add some (perhaps trivial or non-useful) feature to @value{GDBN}, use
6828 the code to predict what a test case would do and write the test case
6829 and verify your prediction, etc. If you are reading code and your eyes
6830 are starting to glaze over, this is a sign you need to use a more active
6833 Once you have a part of @value{GDBN} to start with, you can find more
6834 specifically the part you are looking for by stepping through each
6835 function with the @code{next} command. Do not use @code{step} or you
6836 will quickly get distracted; when the function you are stepping through
6837 calls another function try only to get a big-picture understanding
6838 (perhaps using the comment at the beginning of the function being
6839 called) of what it does. This way you can identify which of the
6840 functions being called by the function you are stepping through is the
6841 one which you are interested in. You may need to examine the data
6842 structures generated at each stage, with reference to the comments in
6843 the header files explaining what the data structures are supposed to
6846 Of course, this same technique can be used if you are just reading the
6847 code, rather than actually stepping through it. The same general
6848 principle applies---when the code you are looking at calls something
6849 else, just try to understand generally what the code being called does,
6850 rather than worrying about all its details.
6852 @cindex command implementation
6853 A good place to start when tracking down some particular area is with
6854 a command which invokes that feature. Suppose you want to know how
6855 single-stepping works. As a @value{GDBN} user, you know that the
6856 @code{step} command invokes single-stepping. The command is invoked
6857 via command tables (see @file{command.h}); by convention the function
6858 which actually performs the command is formed by taking the name of
6859 the command and adding @samp{_command}, or in the case of an
6860 @code{info} subcommand, @samp{_info}. For example, the @code{step}
6861 command invokes the @code{step_command} function and the @code{info
6862 display} command invokes @code{display_info}. When this convention is
6863 not followed, you might have to use @code{grep} or @kbd{M-x
6864 tags-search} in emacs, or run @value{GDBN} on itself and set a
6865 breakpoint in @code{execute_command}.
6867 @cindex @code{bug-gdb} mailing list
6868 If all of the above fail, it may be appropriate to ask for information
6869 on @code{bug-gdb}. But @emph{never} post a generic question like ``I was
6870 wondering if anyone could give me some tips about understanding
6871 @value{GDBN}''---if we had some magic secret we would put it in this manual.
6872 Suggestions for improving the manual are always welcome, of course.
6874 @node Debugging GDB,,,Hints
6876 @section Debugging @value{GDBN} with itself
6877 @cindex debugging @value{GDBN}
6879 If @value{GDBN} is limping on your machine, this is the preferred way to get it
6880 fully functional. Be warned that in some ancient Unix systems, like
6881 Ultrix 4.2, a program can't be running in one process while it is being
6882 debugged in another. Rather than typing the command @kbd{@w{./gdb
6883 ./gdb}}, which works on Suns and such, you can copy @file{gdb} to
6884 @file{gdb2} and then type @kbd{@w{./gdb ./gdb2}}.
6886 When you run @value{GDBN} in the @value{GDBN} source directory, it will read a
6887 @file{.gdbinit} file that sets up some simple things to make debugging
6888 gdb easier. The @code{info} command, when executed without a subcommand
6889 in a @value{GDBN} being debugged by gdb, will pop you back up to the top level
6890 gdb. See @file{.gdbinit} for details.
6892 If you use emacs, you will probably want to do a @code{make TAGS} after
6893 you configure your distribution; this will put the machine dependent
6894 routines for your local machine where they will be accessed first by
6897 Also, make sure that you've either compiled @value{GDBN} with your local cc, or
6898 have run @code{fixincludes} if you are compiling with gcc.
6900 @section Submitting Patches
6902 @cindex submitting patches
6903 Thanks for thinking of offering your changes back to the community of
6904 @value{GDBN} users. In general we like to get well designed enhancements.
6905 Thanks also for checking in advance about the best way to transfer the
6908 The @value{GDBN} maintainers will only install ``cleanly designed'' patches.
6909 This manual summarizes what we believe to be clean design for @value{GDBN}.
6911 If the maintainers don't have time to put the patch in when it arrives,
6912 or if there is any question about a patch, it goes into a large queue
6913 with everyone else's patches and bug reports.
6915 @cindex legal papers for code contributions
6916 The legal issue is that to incorporate substantial changes requires a
6917 copyright assignment from you and/or your employer, granting ownership
6918 of the changes to the Free Software Foundation. You can get the
6919 standard documents for doing this by sending mail to @code{gnu@@gnu.org}
6920 and asking for it. We recommend that people write in "All programs
6921 owned by the Free Software Foundation" as "NAME OF PROGRAM", so that
6922 changes in many programs (not just @value{GDBN}, but GAS, Emacs, GCC,
6924 contributed with only one piece of legalese pushed through the
6925 bureaucracy and filed with the FSF. We can't start merging changes until
6926 this paperwork is received by the FSF (their rules, which we follow
6927 since we maintain it for them).
6929 Technically, the easiest way to receive changes is to receive each
6930 feature as a small context diff or unidiff, suitable for @code{patch}.
6931 Each message sent to me should include the changes to C code and
6932 header files for a single feature, plus @file{ChangeLog} entries for
6933 each directory where files were modified, and diffs for any changes
6934 needed to the manuals (@file{gdb/doc/gdb.texinfo} or
6935 @file{gdb/doc/gdbint.texinfo}). If there are a lot of changes for a
6936 single feature, they can be split down into multiple messages.
6938 In this way, if we read and like the feature, we can add it to the
6939 sources with a single patch command, do some testing, and check it in.
6940 If you leave out the @file{ChangeLog}, we have to write one. If you leave
6941 out the doc, we have to puzzle out what needs documenting. Etc., etc.
6943 The reason to send each change in a separate message is that we will not
6944 install some of the changes. They'll be returned to you with questions
6945 or comments. If we're doing our job correctly, the message back to you
6946 will say what you have to fix in order to make the change acceptable.
6947 The reason to have separate messages for separate features is so that
6948 the acceptable changes can be installed while one or more changes are
6949 being reworked. If multiple features are sent in a single message, we
6950 tend to not put in the effort to sort out the acceptable changes from
6951 the unacceptable, so none of the features get installed until all are
6954 If this sounds painful or authoritarian, well, it is. But we get a lot
6955 of bug reports and a lot of patches, and many of them don't get
6956 installed because we don't have the time to finish the job that the bug
6957 reporter or the contributor could have done. Patches that arrive
6958 complete, working, and well designed, tend to get installed on the day
6959 they arrive. The others go into a queue and get installed as time
6960 permits, which, since the maintainers have many demands to meet, may not
6961 be for quite some time.
6963 Please send patches directly to
6964 @email{gdb-patches@@sources.redhat.com, the @value{GDBN} maintainers}.
6966 @section Obsolete Conditionals
6967 @cindex obsolete code
6969 Fragments of old code in @value{GDBN} sometimes reference or set the following
6970 configuration macros. They should not be used by new code, and old uses
6971 should be removed as those parts of the debugger are otherwise touched.
6974 @item STACK_END_ADDR
6975 This macro used to define where the end of the stack appeared, for use
6976 in interpreting core file formats that don't record this address in the
6977 core file itself. This information is now configured in BFD, and @value{GDBN}
6978 gets the info portably from there. The values in @value{GDBN}'s configuration
6979 files should be moved into BFD configuration files (if needed there),
6980 and deleted from all of @value{GDBN}'s config files.
6982 Any @file{@var{foo}-xdep.c} file that references STACK_END_ADDR
6983 is so old that it has never been converted to use BFD. Now that's old!
6987 @include observer.texi