URL
https://opencores.org/ocsvn/openrisc_2011-10-31/openrisc_2011-10-31/trunk
Subversion Repositories openrisc_2011-10-31
[/] [openrisc/] [trunk/] [gnu-src/] [gdb-7.2/] [gdb/] [doc/] [gdbint.info-1] - Rev 357
Go to most recent revision | Compare with Previous | Blame | View Log
This is gdbint.info, produced by makeinfo version 4.13 from
./gdbint.texinfo.
INFO-DIR-SECTION Software development
START-INFO-DIR-ENTRY
* Gdb-Internals: (gdbint). The GNU debugger's internals.
END-INFO-DIR-ENTRY
Copyright (C) 1990, 1991, 1992, 1993, 1994, 1996, 1998, 1999, 2000,
2001, 2002, 2003, 2004, 2005, 2006, 2008, 2009, 2010 Free Software
Foundation, Inc. Contributed by Cygnus Solutions. Written by John
Gilmore. Second Edition by Stan Shebs.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with no
Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
Texts. A copy of the license is included in the section entitled "GNU
Free Documentation License".
This file documents the internals of the GNU debugger GDB.
Copyright (C) 1990, 1991, 1992, 1993, 1994, 1996, 1998, 1999, 2000,
2001, 2002, 2003, 2004, 2005, 2006, 2008, 2009, 2010 Free Software
Foundation, Inc. Contributed by Cygnus Solutions. Written by John
Gilmore. Second Edition by Stan Shebs.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with no
Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
Texts. A copy of the license is included in the section entitled "GNU
Free Documentation License".
File: gdbint.info, Node: Top, Next: Summary, Up: (dir)
Scope of this Document
**********************
This document documents the internals of the GNU debugger, GDB. It
includes description of GDB's key algorithms and operations, as well as
the mechanisms that adapt GDB to specific hosts and targets.
* Menu:
* Summary::
* Overall Structure::
* Algorithms::
* User Interface::
* libgdb::
* Values::
* Stack Frames::
* Symbol Handling::
* Language Support::
* Host Definition::
* Target Architecture Definition::
* Target Descriptions::
* Target Vector Definition::
* Native Debugging::
* Support Libraries::
* Coding::
* Porting GDB::
* Versions and Branches::
* Start of New Year Procedure::
* Releasing GDB::
* Testsuite::
* Hints::
* GDB Observers:: GDB Currently available observers
* GNU Free Documentation License:: The license for this documentation
* Index::
File: gdbint.info, Node: Summary, Next: Overall Structure, Prev: Top, Up: Top
1 Summary
*********
* Menu:
* Requirements::
* Contributors::
File: gdbint.info, Node: Requirements, Next: Contributors, Up: Summary
1.1 Requirements
================
Before diving into the internals, you should understand the formal
requirements and other expectations for GDB. Although some of these
may seem obvious, there have been proposals for GDB that have run
counter to these requirements.
First of all, GDB is a debugger. It's not designed to be a front
panel for embedded systems. It's not a text editor. It's not a shell.
It's not a programming environment.
GDB is an interactive tool. Although a batch mode is available,
GDB's primary role is to interact with a human programmer.
GDB should be responsive to the user. A programmer hot on the trail
of a nasty bug, and operating under a looming deadline, is going to be
very impatient of everything, including the response time to debugger
commands.
GDB should be relatively permissive, such as for expressions. While
the compiler should be picky (or have the option to be made picky),
since source code lives for a long time usually, the programmer doing
debugging shouldn't be spending time figuring out to mollify the
debugger.
GDB will be called upon to deal with really large programs.
Executable sizes of 50 to 100 megabytes occur regularly, and we've
heard reports of programs approaching 1 gigabyte in size.
GDB should be able to run everywhere. No other debugger is
available for even half as many configurations as GDB supports.
File: gdbint.info, Node: Contributors, Prev: Requirements, Up: Summary
1.2 Contributors
================
The first edition of this document was written by John Gilmore of
Cygnus Solutions. The current second edition was written by Stan Shebs
of Cygnus Solutions, who continues to update the manual.
Over the years, many others have made additions and changes to this
document. This section attempts to record the significant contributors
to that effort. One of the virtues of free software is that everyone is
free to contribute to it; with regret, we cannot actually acknowledge
everyone here.
_Plea:_ This section has only been added relatively recently (four
years after publication of the second edition). Additions to this
section are particularly welcome. If you or your friends (or
enemies, to be evenhanded) have been unfairly omitted from this
list, we would like to add your names!
A document such as this relies on being kept up to date by numerous
small updates by contributing engineers as they make changes to the
code base. The file `ChangeLog' in the GDB distribution approximates a
blow-by-blow account. The most prolific contributors to this important,
but low profile task are Andrew Cagney (responsible for over half the
entries), Daniel Jacobowitz, Mark Kettenis, Jim Blandy and Eli
Zaretskii.
Eli Zaretskii and Daniel Jacobowitz wrote the sections documenting
watchpoints.
Jeremy Bennett updated the sections on initializing a new
architecture and register representation, and added the section on
Frame Interpretation.
File: gdbint.info, Node: Overall Structure, Next: Algorithms, Prev: Summary, Up: Top
2 Overall Structure
*******************
GDB consists of three major subsystems: user interface, symbol handling
(the "symbol side"), and target system handling (the "target side").
The user interface consists of several actual interfaces, plus
supporting code.
The symbol side consists of object file readers, debugging info
interpreters, symbol table management, source language expression
parsing, type and value printing.
The target side consists of execution control, stack frame analysis,
and physical target manipulation.
The target side/symbol side division is not formal, and there are a
number of exceptions. For instance, core file support involves symbolic
elements (the basic core file reader is in BFD) and target elements (it
supplies the contents of memory and the values of registers). Instead,
this division is useful for understanding how the minor subsystems
should fit together.
2.1 The Symbol Side
===================
The symbolic side of GDB can be thought of as "everything you can do in
GDB without having a live program running". For instance, you can look
at the types of variables, and evaluate many kinds of expressions.
2.2 The Target Side
===================
The target side of GDB is the "bits and bytes manipulator". Although
it may make reference to symbolic info here and there, most of the
target side will run with only a stripped executable available--or even
no executable at all, in remote debugging cases.
Operations such as disassembly, stack frame crawls, and register
display, are able to work with no symbolic info at all. In some cases,
such as disassembly, GDB will use symbolic info to present addresses
relative to symbols rather than as raw numbers, but it will work either
way.
2.3 Configurations
==================
"Host" refers to attributes of the system where GDB runs. "Target"
refers to the system where the program being debugged executes. In
most cases they are the same machine, in which case a third type of
"Native" attributes come into play.
Defines and include files needed to build on the host are host
support. Examples are tty support, system defined types, host byte
order, host float format. These are all calculated by `autoconf' when
the debugger is built.
Defines and information needed to handle the target format are target
dependent. Examples are the stack frame format, instruction set,
breakpoint instruction, registers, and how to set up and tear down the
stack to call a function.
Information that is only needed when the host and target are the
same, is native dependent. One example is Unix child process support;
if the host and target are not the same, calling `fork' to start the
target process is a bad idea. The various macros needed for finding the
registers in the `upage', running `ptrace', and such are all in the
native-dependent files.
Another example of native-dependent code is support for features that
are really part of the target environment, but which require `#include'
files that are only available on the host system. Core file handling
and `setjmp' handling are two common cases.
When you want to make GDB work as the traditional native debugger on
a system, you will need to supply both target and native information.
2.4 Source Tree Structure
=========================
The GDB source directory has a mostly flat structure--there are only a
few subdirectories. A file's name usually gives a hint as to what it
does; for example, `stabsread.c' reads stabs, `dwarf2read.c' reads
DWARF 2, etc.
Files that are related to some common task have names that share
common substrings. For example, `*-thread.c' files deal with debugging
threads on various platforms; `*read.c' files deal with reading various
kinds of symbol and object files; `inf*.c' files deal with direct
control of the "inferior program" (GDB parlance for the program being
debugged).
There are several dozens of files in the `*-tdep.c' family. `tdep'
stands for "target-dependent code"--each of these files implements
debug support for a specific target architecture (sparc, mips, etc).
Usually, only one of these will be used in a specific GDB configuration
(sometimes two, closely related).
Similarly, there are many `*-nat.c' files, each one for native
debugging on a specific system (e.g., `sparc-linux-nat.c' is for native
debugging of Sparc machines running the Linux kernel).
The few subdirectories of the source tree are:
`cli'
Code that implements "CLI", the GDB Command-Line Interpreter.
*Note Command Interpreter: User Interface.
`gdbserver'
Code for the GDB remote server.
`gdbtk'
Code for Insight, the GDB TK-based GUI front-end.
`mi'
The "GDB/MI", the GDB Machine Interface interpreter.
`signals'
Target signal translation code.
`tui'
Code for "TUI", the GDB Text-mode full-screen User Interface.
*Note TUI: User Interface.
File: gdbint.info, Node: Algorithms, Next: User Interface, Prev: Overall Structure, Up: Top
3 Algorithms
************
GDB uses a number of debugging-specific algorithms. They are often not
very complicated, but get lost in the thicket of special cases and
real-world issues. This chapter describes the basic algorithms and
mentions some of the specific target definitions that they use.
3.1 Prologue Analysis
=====================
To produce a backtrace and allow the user to manipulate older frames'
variables and arguments, GDB needs to find the base addresses of older
frames, and discover where those frames' registers have been saved.
Since a frame's "callee-saves" registers get saved by younger frames if
and when they're reused, a frame's registers may be scattered
unpredictably across younger frames. This means that changing the
value of a register-allocated variable in an older frame may actually
entail writing to a save slot in some younger frame.
Modern versions of GCC emit Dwarf call frame information ("CFI"),
which describes how to find frame base addresses and saved registers.
But CFI is not always available, so as a fallback GDB uses a technique
called "prologue analysis" to find frame sizes and saved registers. A
prologue analyzer disassembles the function's machine code starting
from its entry point, and looks for instructions that allocate frame
space, save the stack pointer in a frame pointer register, save
registers, and so on. Obviously, this can't be done accurately in
general, but it's tractable to do well enough to be very helpful.
Prologue analysis predates the GNU toolchain's support for CFI; at one
time, prologue analysis was the only mechanism GDB used for stack
unwinding at all, when the function calling conventions didn't specify
a fixed frame layout.
In the olden days, function prologues were generated by hand-written,
target-specific code in GCC, and treated as opaque and untouchable by
optimizers. Looking at this code, it was usually straightforward to
write a prologue analyzer for GDB that would accurately understand all
the prologues GCC would generate. However, over time GCC became more
aggressive about instruction scheduling, and began to understand more
about the semantics of the prologue instructions themselves; in
response, GDB's analyzers became more complex and fragile. Keeping the
prologue analyzers working as GCC (and the instruction sets themselves)
evolved became a substantial task.
To try to address this problem, the code in `prologue-value.h' and
`prologue-value.c' provides a general framework for writing prologue
analyzers that are simpler and more robust than ad-hoc analyzers. When
we analyze a prologue using the prologue-value framework, we're really
doing "abstract interpretation" or "pseudo-evaluation": running the
function's code in simulation, but using conservative approximations of
the values registers and memory would hold when the code actually runs.
For example, if our function starts with the instruction:
addi r1, 42 # add 42 to r1
we don't know exactly what value will be in `r1' after executing
this instruction, but we do know it'll be 42 greater than its original
value.
If we then see an instruction like:
addi r1, 22 # add 22 to r1
we still don't know what `r1's' value is, but again, we can say it
is now 64 greater than its original value.
If the next instruction were:
mov r2, r1 # set r2 to r1's value
then we can say that `r2's' value is now the original value of `r1'
plus 64.
It's common for prologues to save registers on the stack, so we'll
need to track the values of stack frame slots, as well as the
registers. So after an instruction like this:
mov (fp+4), r2
then we'd know that the stack slot four bytes above the frame pointer
holds the original value of `r1' plus 64.
And so on.
Of course, this can only go so far before it gets unreasonable. If
we wanted to be able to say anything about the value of `r1' after the
instruction:
xor r1, r3 # exclusive-or r1 and r3, place result in r1
then things would get pretty complex. But remember, we're just doing
a conservative approximation; if exclusive-or instructions aren't
relevant to prologues, we can just say `r1''s value is now "unknown".
We can ignore things that are too complex, if that loss of information
is acceptable for our application.
So when we say "conservative approximation" here, what we mean is an
approximation that is either accurate, or marked "unknown", but never
inaccurate.
Using this framework, a prologue analyzer is simply an interpreter
for machine code, but one that uses conservative approximations for the
contents of registers and memory instead of actual values. Starting
from the function's entry point, you simulate instructions up to the
current PC, or an instruction that you don't know how to simulate. Now
you can examine the state of the registers and stack slots you've kept
track of.
* To see how large your stack frame is, just check the value of the
stack pointer register; if it's the original value of the SP minus
a constant, then that constant is the stack frame's size. If the
SP's value has been marked as "unknown", then that means the
prologue has done something too complex for us to track, and we
don't know the frame size.
* To see where we've saved the previous frame's registers, we just
search the values we've tracked -- stack slots, usually, but
registers, too, if you want -- for something equal to the
register's original value. If the calling conventions suggest a
standard place to save a given register, then we can check there
first, but really, anything that will get us back the original
value will probably work.
This does take some work. But prologue analyzers aren't
quick-and-simple pattern patching to recognize a few fixed prologue
forms any more; they're big, hairy functions. Along with inferior
function calls, prologue analysis accounts for a substantial portion of
the time needed to stabilize a GDB port. So it's worthwhile to look
for an approach that will be easier to understand and maintain. In the
approach described above:
* It's easier to see that the analyzer is correct: you just see
whether the analyzer properly (albeit conservatively) simulates
the effect of each instruction.
* It's easier to extend the analyzer: you can add support for new
instructions, and know that you haven't broken anything that
wasn't already broken before.
* It's orthogonal: to gather new information, you don't need to
complicate the code for each instruction. As long as your domain
of conservative values is already detailed enough to tell you what
you need, then all the existing instruction simulations are
already gathering the right data for you.
The file `prologue-value.h' contains detailed comments explaining
the framework and how to use it.
3.2 Breakpoint Handling
=======================
In general, a breakpoint is a user-designated location in the program
where the user wants to regain control if program execution ever reaches
that location.
There are two main ways to implement breakpoints; either as
"hardware" breakpoints or as "software" breakpoints.
Hardware breakpoints are sometimes available as a builtin debugging
features with some chips. Typically these work by having dedicated
register into which the breakpoint address may be stored. If the PC
(shorthand for "program counter") ever matches a value in a breakpoint
registers, the CPU raises an exception and reports it to GDB.
Another possibility is when an emulator is in use; many emulators
include circuitry that watches the address lines coming out from the
processor, and force it to stop if the address matches a breakpoint's
address.
A third possibility is that the target already has the ability to do
breakpoints somehow; for instance, a ROM monitor may do its own
software breakpoints. So although these are not literally "hardware
breakpoints", from GDB's point of view they work the same; GDB need not
do anything more than set the breakpoint and wait for something to
happen.
Since they depend on hardware resources, hardware breakpoints may be
limited in number; when the user asks for more, GDB will start trying
to set software breakpoints. (On some architectures, notably the
32-bit x86 platforms, GDB cannot always know whether there's enough
hardware resources to insert all the hardware breakpoints and
watchpoints. On those platforms, GDB prints an error message only when
the program being debugged is continued.)
Software breakpoints require GDB to do somewhat more work. The
basic theory is that GDB will replace a program instruction with a
trap, illegal divide, or some other instruction that will cause an
exception, and then when it's encountered, GDB will take the exception
and stop the program. When the user says to continue, GDB will restore
the original instruction, single-step, re-insert the trap, and continue
on.
Since it literally overwrites the program being tested, the program
area must be writable, so this technique won't work on programs in ROM.
It can also distort the behavior of programs that examine themselves,
although such a situation would be highly unusual.
Also, the software breakpoint instruction should be the smallest
size of instruction, so it doesn't overwrite an instruction that might
be a jump target, and cause disaster when the program jumps into the
middle of the breakpoint instruction. (Strictly speaking, the
breakpoint must be no larger than the smallest interval between
instructions that may be jump targets; perhaps there is an architecture
where only even-numbered instructions may jumped to.) Note that it's
possible for an instruction set not to have any instructions usable for
a software breakpoint, although in practice only the ARC has failed to
define such an instruction.
Basic breakpoint object handling is in `breakpoint.c'. However,
much of the interesting breakpoint action is in `infrun.c'.
`target_remove_breakpoint (BP_TGT)'
`target_insert_breakpoint (BP_TGT)'
Insert or remove a software breakpoint at address
`BP_TGT->placed_address'. Returns zero for success, non-zero for
failure. On input, BP_TGT contains the address of the breakpoint,
and is otherwise initialized to zero. The fields of the `struct
bp_target_info' pointed to by BP_TGT are updated to contain other
information about the breakpoint on output. The field
`placed_address' may be updated if the breakpoint was placed at a
related address; the field `shadow_contents' contains the real
contents of the bytes where the breakpoint has been inserted, if
reading memory would return the breakpoint instead of the
underlying memory; the field `shadow_len' is the length of memory
cached in `shadow_contents', if any; and the field `placed_size'
is optionally set and used by the target, if it could differ from
`shadow_len'.
For example, the remote target `Z0' packet does not require
shadowing memory, so `shadow_len' is left at zero. However, the
length reported by `gdbarch_breakpoint_from_pc' is cached in
`placed_size', so that a matching `z0' packet can be used to
remove the breakpoint.
`target_remove_hw_breakpoint (BP_TGT)'
`target_insert_hw_breakpoint (BP_TGT)'
Insert or remove a hardware-assisted breakpoint at address
`BP_TGT->placed_address'. Returns zero for success, non-zero for
failure. See `target_insert_breakpoint' for a description of the
`struct bp_target_info' pointed to by BP_TGT; the
`shadow_contents' and `shadow_len' members are not used for
hardware breakpoints, but `placed_size' may be.
3.3 Single Stepping
===================
3.4 Signal Handling
===================
3.5 Thread Handling
===================
3.6 Inferior Function Calls
===========================
3.7 Longjmp Support
===================
GDB has support for figuring out that the target is doing a `longjmp'
and for stopping at the target of the jump, if we are stepping. This
is done with a few specialized internal breakpoints, which are visible
in the output of the `maint info breakpoint' command.
To make this work, you need to define a function called
`gdbarch_get_longjmp_target', which will examine the `jmp_buf'
structure and extract the `longjmp' target address. Since `jmp_buf' is
target specific and typically defined in a target header not available
to GDB, you will need to determine the offset of the PC manually and
return that; many targets define a `jb_pc_offset' field in the tdep
structure to save the value once calculated.
3.8 Watchpoints
===============
Watchpoints are a special kind of breakpoints (*note breakpoints:
Algorithms.) which break when data is accessed rather than when some
instruction is executed. When you have data which changes without your
knowing what code does that, watchpoints are the silver bullet to hunt
down and kill such bugs.
Watchpoints can be either hardware-assisted or not; the latter type
is known as "software watchpoints." GDB always uses hardware-assisted
watchpoints if they are available, and falls back on software
watchpoints otherwise. Typical situations where GDB will use software
watchpoints are:
* The watched memory region is too large for the underlying hardware
watchpoint support. For example, each x86 debug register can
watch up to 4 bytes of memory, so trying to watch data structures
whose size is more than 16 bytes will cause GDB to use software
watchpoints.
* The value of the expression to be watched depends on data held in
registers (as opposed to memory).
* Too many different watchpoints requested. (On some architectures,
this situation is impossible to detect until the debugged program
is resumed.) Note that x86 debug registers are used both for
hardware breakpoints and for watchpoints, so setting too many
hardware breakpoints might cause watchpoint insertion to fail.
* No hardware-assisted watchpoints provided by the target
implementation.
Software watchpoints are very slow, since GDB needs to single-step
the program being debugged and test the value of the watched
expression(s) after each instruction. The rest of this section is
mostly irrelevant for software watchpoints.
When the inferior stops, GDB tries to establish, among other
possible reasons, whether it stopped due to a watchpoint being hit. It
first uses `STOPPED_BY_WATCHPOINT' to see if any watchpoint was hit.
If not, all watchpoint checking is skipped.
Then GDB calls `target_stopped_data_address' exactly once. This
method returns the address of the watchpoint which triggered, if the
target can determine it. If the triggered address is available, GDB
compares the address returned by this method with each watched memory
address in each active watchpoint. For data-read and data-access
watchpoints, GDB announces every watchpoint that watches the triggered
address as being hit. For this reason, data-read and data-access
watchpoints _require_ that the triggered address be available; if not,
read and access watchpoints will never be considered hit. For
data-write watchpoints, if the triggered address is available, GDB
considers only those watchpoints which match that address; otherwise,
GDB considers all data-write watchpoints. For each data-write
watchpoint that GDB considers, it evaluates the expression whose value
is being watched, and tests whether the watched value has changed.
Watchpoints whose watched values have changed are announced as hit.
GDB uses several macros and primitives to support hardware
watchpoints:
`TARGET_CAN_USE_HARDWARE_WATCHPOINT (TYPE, COUNT, OTHER)'
Return the number of hardware watchpoints of type TYPE that are
possible to be set. The value is positive if COUNT watchpoints of
this type can be set, zero if setting watchpoints of this type is
not supported, and negative if COUNT is more than the maximum
number of watchpoints of type TYPE that can be set. OTHER is
non-zero if other types of watchpoints are currently enabled (there
are architectures which cannot set watchpoints of different types
at the same time).
`TARGET_REGION_OK_FOR_HW_WATCHPOINT (ADDR, LEN)'
Return non-zero if hardware watchpoints can be used to watch a
region whose address is ADDR and whose length in bytes is LEN.
`target_insert_watchpoint (ADDR, LEN, TYPE)'
`target_remove_watchpoint (ADDR, LEN, TYPE)'
Insert or remove a hardware watchpoint starting at ADDR, for LEN
bytes. TYPE is the watchpoint type, one of the possible values of
the enumerated data type `target_hw_bp_type', defined by
`breakpoint.h' as follows:
enum target_hw_bp_type
{
hw_write = 0, /* Common (write) HW watchpoint */
hw_read = 1, /* Read HW watchpoint */
hw_access = 2, /* Access (read or write) HW watchpoint */
hw_execute = 3 /* Execute HW breakpoint */
};
These two macros should return 0 for success, non-zero for failure.
`target_stopped_data_address (ADDR_P)'
If the inferior has some watchpoint that triggered, place the
address associated with the watchpoint at the location pointed to
by ADDR_P and return non-zero. Otherwise, return zero. This is
required for data-read and data-access watchpoints. It is not
required for data-write watchpoints, but GDB uses it to improve
handling of those also.
GDB will only call this method once per watchpoint stop,
immediately after calling `STOPPED_BY_WATCHPOINT'. If the
target's watchpoint indication is sticky, i.e., stays set after
resuming, this method should clear it. For instance, the x86 debug
control register has sticky triggered flags.
`target_watchpoint_addr_within_range (TARGET, ADDR, START, LENGTH)'
Check whether ADDR (as returned by `target_stopped_data_address')
lies within the hardware-defined watchpoint region described by
START and LENGTH. This only needs to be provided if the
granularity of a watchpoint is greater than one byte, i.e., if the
watchpoint can also trigger on nearby addresses outside of the
watched region.
`HAVE_STEPPABLE_WATCHPOINT'
If defined to a non-zero value, it is not necessary to disable a
watchpoint to step over it. Like
`gdbarch_have_nonsteppable_watchpoint', this is usually set when
watchpoints trigger at the instruction which will perform an
interesting read or write. It should be set if there is a
temporary disable bit which allows the processor to step over the
interesting instruction without raising the watchpoint exception
again.
`int gdbarch_have_nonsteppable_watchpoint (GDBARCH)'
If it returns a non-zero value, GDB should disable a watchpoint to
step the inferior over it. This is usually set when watchpoints
trigger at the instruction which will perform an interesting read
or write.
`HAVE_CONTINUABLE_WATCHPOINT'
If defined to a non-zero value, it is possible to continue the
inferior after a watchpoint has been hit. This is usually set
when watchpoints trigger at the instruction following an
interesting read or write.
`STOPPED_BY_WATCHPOINT (WAIT_STATUS)'
Return non-zero if stopped by a watchpoint. WAIT_STATUS is of the
type `struct target_waitstatus', defined by `target.h'. Normally,
this macro is defined to invoke the function pointed to by the
`to_stopped_by_watchpoint' member of the structure (of the type
`target_ops', defined on `target.h') that describes the
target-specific operations; `to_stopped_by_watchpoint' ignores the
WAIT_STATUS argument.
GDB does not require the non-zero value returned by
`STOPPED_BY_WATCHPOINT' to be 100% correct, so if a target cannot
determine for sure whether the inferior stopped due to a
watchpoint, it could return non-zero "just in case".
3.8.1 Watchpoints and Threads
-----------------------------
GDB only supports process-wide watchpoints, which trigger in all
threads. GDB uses the thread ID to make watchpoints act as if they
were thread-specific, but it cannot set hardware watchpoints that only
trigger in a specific thread. Therefore, even if the target supports
threads, per-thread debug registers, and watchpoints which only affect
a single thread, it should set the per-thread debug registers for all
threads to the same value. On GNU/Linux native targets, this is
accomplished by using `ALL_LWPS' in `target_insert_watchpoint' and
`target_remove_watchpoint' and by using `linux_set_new_thread' to
register a handler for newly created threads.
GDB's GNU/Linux support only reports a single event at a time,
although multiple events can trigger simultaneously for multi-threaded
programs. When multiple events occur, `linux-nat.c' queues subsequent
events and returns them the next time the program is resumed. This
means that `STOPPED_BY_WATCHPOINT' and `target_stopped_data_address'
only need to consult the current thread's state--the thread indicated
by `inferior_ptid'. If two threads have hit watchpoints
simultaneously, those routines will be called a second time for the
second thread.
3.8.2 x86 Watchpoints
---------------------
The 32-bit Intel x86 (a.k.a. ia32) processors feature special debug
registers designed to facilitate debugging. GDB provides a generic
library of functions that x86-based ports can use to implement support
for watchpoints and hardware-assisted breakpoints. This subsection
documents the x86 watchpoint facilities in GDB.
(At present, the library functions read and write debug registers
directly, and are thus only available for native configurations.)
To use the generic x86 watchpoint support, a port should do the
following:
* Define the macro `I386_USE_GENERIC_WATCHPOINTS' somewhere in the
target-dependent headers.
* Include the `config/i386/nm-i386.h' header file _after_ defining
`I386_USE_GENERIC_WATCHPOINTS'.
* Add `i386-nat.o' to the value of the Make variable `NATDEPFILES'
(*note NATDEPFILES: Native Debugging.).
* Provide implementations for the `I386_DR_LOW_*' macros described
below. Typically, each macro should call a target-specific
function which does the real work.
The x86 watchpoint support works by maintaining mirror images of the
debug registers. Values are copied between the mirror images and the
real debug registers via a set of macros which each target needs to
provide:
`I386_DR_LOW_SET_CONTROL (VAL)'
Set the Debug Control (DR7) register to the value VAL.
`I386_DR_LOW_SET_ADDR (IDX, ADDR)'
Put the address ADDR into the debug register number IDX.
`I386_DR_LOW_RESET_ADDR (IDX)'
Reset (i.e. zero out) the address stored in the debug register
number IDX.
`I386_DR_LOW_GET_STATUS'
Return the value of the Debug Status (DR6) register. This value is
used immediately after it is returned by `I386_DR_LOW_GET_STATUS',
so as to support per-thread status register values.
For each one of the 4 debug registers (whose indices are from 0 to 3)
that store addresses, a reference count is maintained by GDB, to allow
sharing of debug registers by several watchpoints. This allows users
to define several watchpoints that watch the same expression, but with
different conditions and/or commands, without wasting debug registers
which are in short supply. GDB maintains the reference counts
internally, targets don't have to do anything to use this feature.
The x86 debug registers can each watch a region that is 1, 2, or 4
bytes long. The ia32 architecture requires that each watched region be
appropriately aligned: 2-byte region on 2-byte boundary, 4-byte region
on 4-byte boundary. However, the x86 watchpoint support in GDB can
watch unaligned regions and regions larger than 4 bytes (up to 16
bytes) by allocating several debug registers to watch a single region.
This allocation of several registers per a watched region is also done
automatically without target code intervention.
The generic x86 watchpoint support provides the following API for the
GDB's application code:
`i386_region_ok_for_watchpoint (ADDR, LEN)'
The macro `TARGET_REGION_OK_FOR_HW_WATCHPOINT' is set to call this
function. It counts the number of debug registers required to
watch a given region, and returns a non-zero value if that number
is less than 4, the number of debug registers available to x86
processors.
`i386_stopped_data_address (ADDR_P)'
The target function `target_stopped_data_address' is set to call
this function. This function examines the breakpoint condition
bits in the DR6 Debug Status register, as returned by the
`I386_DR_LOW_GET_STATUS' macro, and returns the address associated
with the first bit that is set in DR6.
`i386_stopped_by_watchpoint (void)'
The macro `STOPPED_BY_WATCHPOINT' is set to call this function.
The argument passed to `STOPPED_BY_WATCHPOINT' is ignored. This
function examines the breakpoint condition bits in the DR6 Debug
Status register, as returned by the `I386_DR_LOW_GET_STATUS'
macro, and returns true if any bit is set. Otherwise, false is
returned.
`i386_insert_watchpoint (ADDR, LEN, TYPE)'
`i386_remove_watchpoint (ADDR, LEN, TYPE)'
Insert or remove a watchpoint. The macros
`target_insert_watchpoint' and `target_remove_watchpoint' are set
to call these functions. `i386_insert_watchpoint' first looks for
a debug register which is already set to watch the same region for
the same access types; if found, it just increments the reference
count of that debug register, thus implementing debug register
sharing between watchpoints. If no such register is found, the
function looks for a vacant debug register, sets its mirrored
value to ADDR, sets the mirrored value of DR7 Debug Control
register as appropriate for the LEN and TYPE parameters, and then
passes the new values of the debug register and DR7 to the
inferior by calling `I386_DR_LOW_SET_ADDR' and
`I386_DR_LOW_SET_CONTROL'. If more than one debug register is
required to cover the given region, the above process is repeated
for each debug register.
`i386_remove_watchpoint' does the opposite: it resets the address
in the mirrored value of the debug register and its read/write and
length bits in the mirrored value of DR7, then passes these new
values to the inferior via `I386_DR_LOW_RESET_ADDR' and
`I386_DR_LOW_SET_CONTROL'. If a register is shared by several
watchpoints, each time a `i386_remove_watchpoint' is called, it
decrements the reference count, and only calls
`I386_DR_LOW_RESET_ADDR' and `I386_DR_LOW_SET_CONTROL' when the
count goes to zero.
`i386_insert_hw_breakpoint (BP_TGT)'
`i386_remove_hw_breakpoint (BP_TGT)'
These functions insert and remove hardware-assisted breakpoints.
The macros `target_insert_hw_breakpoint' and
`target_remove_hw_breakpoint' are set to call these functions.
The argument is a `struct bp_target_info *', as described in the
documentation for `target_insert_breakpoint'. These functions
work like `i386_insert_watchpoint' and `i386_remove_watchpoint',
respectively, except that they set up the debug registers to watch
instruction execution, and each hardware-assisted breakpoint
always requires exactly one debug register.
`i386_cleanup_dregs (void)'
This function clears all the reference counts, addresses, and
control bits in the mirror images of the debug registers. It
doesn't affect the actual debug registers in the inferior process.
*Notes:*
1. x86 processors support setting watchpoints on I/O reads or writes.
However, since no target supports this (as of March 2001), and
since `enum target_hw_bp_type' doesn't even have an enumeration
for I/O watchpoints, this feature is not yet available to GDB
running on x86.
2. x86 processors can enable watchpoints locally, for the current task
only, or globally, for all the tasks. For each debug register,
there's a bit in the DR7 Debug Control register that determines
whether the associated address is watched locally or globally. The
current implementation of x86 watchpoint support in GDB always
sets watchpoints to be locally enabled, since global watchpoints
might interfere with the underlying OS and are probably
unavailable in many platforms.
3.9 Checkpoints
===============
In the abstract, a checkpoint is a point in the execution history of
the program, which the user may wish to return to at some later time.
Internally, a checkpoint is a saved copy of the program state,
including whatever information is required in order to restore the
program to that state at a later time. This can be expected to include
the state of registers and memory, and may include external state such
as the state of open files and devices.
There are a number of ways in which checkpoints may be implemented
in gdb, e.g. as corefiles, as forked processes, and as some opaque
method implemented on the target side.
A corefile can be used to save an image of target memory and register
state, which can in principle be restored later -- but corefiles do not
typically include information about external entities such as open
files. Currently this method is not implemented in gdb.
A forked process can save the state of user memory and registers, as
well as some subset of external (kernel) state. This method is used to
implement checkpoints on Linux, and in principle might be used on other
systems.
Some targets, e.g. simulators, might have their own built-in method
for saving checkpoints, and gdb might be able to take advantage of that
capability without necessarily knowing any details of how it is done.
3.10 Observing changes in GDB internals
=======================================
In order to function properly, several modules need to be notified when
some changes occur in the GDB internals. Traditionally, these modules
have relied on several paradigms, the most common ones being hooks and
gdb-events. Unfortunately, none of these paradigms was versatile
enough to become the standard notification mechanism in GDB. The fact
that they only supported one "client" was also a strong limitation.
A new paradigm, based on the Observer pattern of the `Design
Patterns' book, has therefore been implemented. The goal was to provide
a new interface overcoming the issues with the notification mechanisms
previously available. This new interface needed to be strongly typed,
easy to extend, and versatile enough to be used as the standard
interface when adding new notifications.
See *note GDB Observers:: for a brief description of the observers
currently implemented in GDB. The rationale for the current
implementation is also briefly discussed.
File: gdbint.info, Node: User Interface, Next: libgdb, Prev: Algorithms, Up: Top
4 User Interface
****************
GDB has several user interfaces, of which the traditional command-line
interface is perhaps the most familiar.
4.1 Command Interpreter
=======================
The command interpreter in GDB is fairly simple. It is designed to
allow for the set of commands to be augmented dynamically, and also has
a recursive subcommand capability, where the first argument to a
command may itself direct a lookup on a different command list.
For instance, the `set' command just starts a lookup on the
`setlist' command list, while `set thread' recurses to the
`set_thread_cmd_list'.
To add commands in general, use `add_cmd'. `add_com' adds to the
main command list, and should be used for those commands. The usual
place to add commands is in the `_initialize_XYZ' routines at the ends
of most source files.
To add paired `set' and `show' commands, use `add_setshow_cmd' or
`add_setshow_cmd_full'. The former is a slightly simpler interface
which is useful when you don't need to further modify the new command
structures, while the latter returns the new command structures for
manipulation.
Before removing commands from the command set it is a good idea to
deprecate them for some time. Use `deprecate_cmd' on commands or
aliases to set the deprecated flag. `deprecate_cmd' takes a `struct
cmd_list_element' as it's first argument. You can use the return value
from `add_com' or `add_cmd' to deprecate the command immediately after
it is created.
The first time a command is used the user will be warned and offered
a replacement (if one exists). Note that the replacement string passed
to `deprecate_cmd' should be the full name of the command, i.e., the
entire string the user should type at the command line.
4.2 UI-Independent Output--the `ui_out' Functions
=================================================
The `ui_out' functions present an abstraction level for the GDB output
code. They hide the specifics of different user interfaces supported
by GDB, and thus free the programmer from the need to write several
versions of the same code, one each for every UI, to produce output.
4.2.1 Overview and Terminology
------------------------------
In general, execution of each GDB command produces some sort of output,
and can even generate an input request.
Output can be generated for the following purposes:
* to display a _result_ of an operation;
* to convey _info_ or produce side-effects of a requested operation;
* to provide a _notification_ of an asynchronous event (including
progress indication of a prolonged asynchronous operation);
* to display _error messages_ (including warnings);
* to show _debug data_;
* to _query_ or prompt a user for input (a special case).
This section mainly concentrates on how to build result output,
although some of it also applies to other kinds of output.
Generation of output that displays the results of an operation
involves one or more of the following:
* output of the actual data
* formatting the output as appropriate for console output, to make it
easily readable by humans
* machine oriented formatting-a more terse formatting to allow for
easy parsing by programs which read GDB's output
* annotation, whose purpose is to help legacy GUIs to identify
interesting parts in the output
The `ui_out' routines take care of the first three aspects.
Annotations are provided by separate annotation routines. Note that use
of annotations for an interface between a GUI and GDB is deprecated.
Output can be in the form of a single item, which we call a "field";
a "list" consisting of identical fields; a "tuple" consisting of
non-identical fields; or a "table", which is a tuple consisting of a
header and a body. In a BNF-like form:
`<table> ==>'
`<header> <body>'
`<header> ==>'
`{ <column> }'
`<column> ==>'
`<width> <alignment> <title>'
`<body> ==>'
`{<row>}'
4.2.2 General Conventions
-------------------------
Most `ui_out' routines are of type `void', the exceptions are
`ui_out_stream_new' (which returns a pointer to the newly created
object) and the `make_cleanup' routines.
The first parameter is always the `ui_out' vector object, a pointer
to a `struct ui_out'.
The FORMAT parameter is like in `printf' family of functions. When
it is present, there must also be a variable list of arguments
sufficient used to satisfy the `%' specifiers in the supplied format.
When a character string argument is not used in a `ui_out' function
call, a `NULL' pointer has to be supplied instead.
4.2.3 Table, Tuple and List Functions
-------------------------------------
This section introduces `ui_out' routines for building lists, tuples
and tables. The routines to output the actual data items (fields) are
presented in the next section.
To recap: A "tuple" is a sequence of "fields", each field containing
information about an object; a "list" is a sequence of fields where
each field describes an identical object.
Use the "table" functions when your output consists of a list of
rows (tuples) and the console output should include a heading. Use this
even when you are listing just one object but you still want the header.
Tables can not be nested. Tuples and lists can be nested up to a
maximum of five levels.
The overall structure of the table output code is something like
this:
ui_out_table_begin
ui_out_table_header
...
ui_out_table_body
ui_out_tuple_begin
ui_out_field_*
...
ui_out_tuple_end
...
ui_out_table_end
Here is the description of table-, tuple- and list-related `ui_out'
functions:
-- Function: void ui_out_table_begin (struct ui_out *UIOUT, int
NBROFCOLS, int NR_ROWS, const char *TBLID)
The function `ui_out_table_begin' marks the beginning of the output
of a table. It should always be called before any other `ui_out'
function for a given table. NBROFCOLS is the number of columns in
the table. NR_ROWS is the number of rows in the table. TBLID is
an optional string identifying the table. The string pointed to
by TBLID is copied by the implementation of `ui_out_table_begin',
so the application can free the string if it was `malloc'ed.
The companion function `ui_out_table_end', described below, marks
the end of the table's output.
-- Function: void ui_out_table_header (struct ui_out *UIOUT, int
WIDTH, enum ui_align ALIGNMENT, const char *COLHDR)
`ui_out_table_header' provides the header information for a single
table column. You call this function several times, one each for
every column of the table, after `ui_out_table_begin', but before
`ui_out_table_body'.
The value of WIDTH gives the column width in characters. The
value of ALIGNMENT is one of `left', `center', and `right', and it
specifies how to align the header: left-justify, center, or
right-justify it. COLHDR points to a string that specifies the
column header; the implementation copies that string, so column
header strings in `malloc'ed storage can be freed after the call.
-- Function: void ui_out_table_body (struct ui_out *UIOUT)
This function delimits the table header from the table body.
-- Function: void ui_out_table_end (struct ui_out *UIOUT)
This function signals the end of a table's output. It should be
called after the table body has been produced by the list and
field output functions.
There should be exactly one call to `ui_out_table_end' for each
call to `ui_out_table_begin', otherwise the `ui_out' functions
will signal an internal error.
The output of the tuples that represent the table rows must follow
the call to `ui_out_table_body' and precede the call to
`ui_out_table_end'. You build a tuple by calling `ui_out_tuple_begin'
and `ui_out_tuple_end', with suitable calls to functions which actually
output fields between them.
-- Function: void ui_out_tuple_begin (struct ui_out *UIOUT, const char
*ID)
This function marks the beginning of a tuple output. ID points to
an optional string that identifies the tuple; it is copied by the
implementation, and so strings in `malloc'ed storage can be freed
after the call.
-- Function: void ui_out_tuple_end (struct ui_out *UIOUT)
This function signals an end of a tuple output. There should be
exactly one call to `ui_out_tuple_end' for each call to
`ui_out_tuple_begin', otherwise an internal GDB error will be
signaled.
-- Function: struct cleanup * make_cleanup_ui_out_tuple_begin_end
(struct ui_out *UIOUT, const char *ID)
This function first opens the tuple and then establishes a cleanup
(*note Cleanups: Coding.) to close the tuple. It provides a
convenient and correct implementation of the non-portable(1) code
sequence:
struct cleanup *old_cleanup;
ui_out_tuple_begin (uiout, "...");
old_cleanup = make_cleanup ((void(*)(void *)) ui_out_tuple_end,
uiout);
-- Function: void ui_out_list_begin (struct ui_out *UIOUT, const char
*ID)
This function marks the beginning of a list output. ID points to
an optional string that identifies the list; it is copied by the
implementation, and so strings in `malloc'ed storage can be freed
after the call.
-- Function: void ui_out_list_end (struct ui_out *UIOUT)
This function signals an end of a list output. There should be
exactly one call to `ui_out_list_end' for each call to
`ui_out_list_begin', otherwise an internal GDB error will be
signaled.
-- Function: struct cleanup * make_cleanup_ui_out_list_begin_end
(struct ui_out *UIOUT, const char *ID)
Similar to `make_cleanup_ui_out_tuple_begin_end', this function
opens a list and then establishes cleanup (*note Cleanups: Coding.)
that will close the list.
4.2.4 Item Output Functions
---------------------------
The functions described below produce output for the actual data items,
or fields, which contain information about the object.
Choose the appropriate function accordingly to your particular needs.
-- Function: void ui_out_field_fmt (struct ui_out *UIOUT, char
*FLDNAME, char *FORMAT, ...)
This is the most general output function. It produces the
representation of the data in the variable-length argument list
according to formatting specifications in FORMAT, a `printf'-like
format string. The optional argument FLDNAME supplies the name of
the field. The data items themselves are supplied as additional
arguments after FORMAT.
This generic function should be used only when it is not possible
to use one of the specialized versions (see below).
-- Function: void ui_out_field_int (struct ui_out *UIOUT, const char
*FLDNAME, int VALUE)
This function outputs a value of an `int' variable. It uses the
`"%d"' output conversion specification. FLDNAME specifies the
name of the field.
-- Function: void ui_out_field_fmt_int (struct ui_out *UIOUT, int
WIDTH, enum ui_align ALIGNMENT, const char *FLDNAME, int
VALUE)
This function outputs a value of an `int' variable. It differs
from `ui_out_field_int' in that the caller specifies the desired
WIDTH and ALIGNMENT of the output. FLDNAME specifies the name of
the field.
-- Function: void ui_out_field_core_addr (struct ui_out *UIOUT, const
char *FLDNAME, struct gdbarch *GDBARCH, CORE_ADDR ADDRESS)
This function outputs an address as appropriate for GDBARCH.
-- Function: void ui_out_field_string (struct ui_out *UIOUT, const
char *FLDNAME, const char *STRING)
This function outputs a string using the `"%s"' conversion
specification.
Sometimes, there's a need to compose your output piece by piece using
functions that operate on a stream, such as `value_print' or
`fprintf_symbol_filtered'. These functions accept an argument of the
type `struct ui_file *', a pointer to a `ui_file' object used to store
the data stream used for the output. When you use one of these
functions, you need a way to pass their results stored in a `ui_file'
object to the `ui_out' functions. To this end, you first create a
`ui_stream' object by calling `ui_out_stream_new', pass the `stream'
member of that `ui_stream' object to `value_print' and similar
functions, and finally call `ui_out_field_stream' to output the field
you constructed. When the `ui_stream' object is no longer needed, you
should destroy it and free its memory by calling `ui_out_stream_delete'.
-- Function: struct ui_stream * ui_out_stream_new (struct ui_out
*UIOUT)
This function creates a new `ui_stream' object which uses the same
output methods as the `ui_out' object whose pointer is passed in
UIOUT. It returns a pointer to the newly created `ui_stream'
object.
-- Function: void ui_out_stream_delete (struct ui_stream *STREAMBUF)
This functions destroys a `ui_stream' object specified by
STREAMBUF.
-- Function: void ui_out_field_stream (struct ui_out *UIOUT, const
char *FIELDNAME, struct ui_stream *STREAMBUF)
This function consumes all the data accumulated in
`streambuf->stream' and outputs it like `ui_out_field_string'
does. After a call to `ui_out_field_stream', the accumulated data
no longer exists, but the stream is still valid and may be used
for producing more fields.
*Important:* If there is any chance that your code could bail out
before completing output generation and reaching the point where
`ui_out_stream_delete' is called, it is necessary to set up a cleanup,
to avoid leaking memory and other resources. Here's a skeleton code to
do that:
struct ui_stream *mybuf = ui_out_stream_new (uiout);
struct cleanup *old = make_cleanup (ui_out_stream_delete, mybuf);
...
do_cleanups (old);
If the function already has the old cleanup chain set (for other
kinds of cleanups), you just have to add your cleanup to it:
mybuf = ui_out_stream_new (uiout);
make_cleanup (ui_out_stream_delete, mybuf);
Note that with cleanups in place, you should not call
`ui_out_stream_delete' directly, or you would attempt to free the same
buffer twice.
4.2.5 Utility Output Functions
------------------------------
-- Function: void ui_out_field_skip (struct ui_out *UIOUT, const char
*FLDNAME)
This function skips a field in a table. Use it if you have to
leave an empty field without disrupting the table alignment. The
argument FLDNAME specifies a name for the (missing) filed.
-- Function: void ui_out_text (struct ui_out *UIOUT, const char
*STRING)
This function outputs the text in STRING in a way that makes it
easy to be read by humans. For example, the console
implementation of this method filters the text through a built-in
pager, to prevent it from scrolling off the visible portion of the
screen.
Use this function for printing relatively long chunks of text
around the actual field data: the text it produces is not aligned
according to the table's format. Use `ui_out_field_string' to
output a string field, and use `ui_out_message', described below,
to output short messages.
-- Function: void ui_out_spaces (struct ui_out *UIOUT, int NSPACES)
This function outputs NSPACES spaces. It is handy to align the
text produced by `ui_out_text' with the rest of the table or list.
-- Function: void ui_out_message (struct ui_out *UIOUT, int VERBOSITY,
const char *FORMAT, ...)
This function produces a formatted message, provided that the
current verbosity level is at least as large as given by
VERBOSITY. The current verbosity level is specified by the user
with the `set verbositylevel' command.(2)
-- Function: void ui_out_wrap_hint (struct ui_out *UIOUT, char *INDENT)
This function gives the console output filter (a paging filter) a
hint of where to break lines which are too long. Ignored for all
other output consumers. INDENT, if non-`NULL', is the string to
be printed to indent the wrapped text on the next line; it must
remain accessible until the next call to `ui_out_wrap_hint', or
until an explicit newline is produced by one of the other
functions. If INDENT is `NULL', the wrapped text will not be
indented.
-- Function: void ui_out_flush (struct ui_out *UIOUT)
This function flushes whatever output has been accumulated so far,
if the UI buffers output.
4.2.6 Examples of Use of `ui_out' functions
-------------------------------------------
This section gives some practical examples of using the `ui_out'
functions to generalize the old console-oriented code in GDB. The
examples all come from functions defined on the `breakpoints.c' file.
This example, from the `breakpoint_1' function, shows how to produce
a table.
The original code was:
if (!found_a_breakpoint++)
{
annotate_breakpoints_headers ();
annotate_field (0);
printf_filtered ("Num ");
annotate_field (1);
printf_filtered ("Type ");
annotate_field (2);
printf_filtered ("Disp ");
annotate_field (3);
printf_filtered ("Enb ");
if (addressprint)
{
annotate_field (4);
printf_filtered ("Address ");
}
annotate_field (5);
printf_filtered ("What\n");
annotate_breakpoints_table ();
}
Here's the new version:
nr_printable_breakpoints = ...;
if (addressprint)
ui_out_table_begin (ui, 6, nr_printable_breakpoints, "BreakpointTable");
else
ui_out_table_begin (ui, 5, nr_printable_breakpoints, "BreakpointTable");
if (nr_printable_breakpoints > 0)
annotate_breakpoints_headers ();
if (nr_printable_breakpoints > 0)
annotate_field (0);
ui_out_table_header (uiout, 3, ui_left, "number", "Num"); /* 1 */
if (nr_printable_breakpoints > 0)
annotate_field (1);
ui_out_table_header (uiout, 14, ui_left, "type", "Type"); /* 2 */
if (nr_printable_breakpoints > 0)
annotate_field (2);
ui_out_table_header (uiout, 4, ui_left, "disp", "Disp"); /* 3 */
if (nr_printable_breakpoints > 0)
annotate_field (3);
ui_out_table_header (uiout, 3, ui_left, "enabled", "Enb"); /* 4 */
if (addressprint)
{
if (nr_printable_breakpoints > 0)
annotate_field (4);
if (print_address_bits <= 32)
ui_out_table_header (uiout, 10, ui_left, "addr", "Address");/* 5 */
else
ui_out_table_header (uiout, 18, ui_left, "addr", "Address");/* 5 */
}
if (nr_printable_breakpoints > 0)
annotate_field (5);
ui_out_table_header (uiout, 40, ui_noalign, "what", "What"); /* 6 */
ui_out_table_body (uiout);
if (nr_printable_breakpoints > 0)
annotate_breakpoints_table ();
This example, from the `print_one_breakpoint' function, shows how to
produce the actual data for the table whose structure was defined in
the above example. The original code was:
annotate_record ();
annotate_field (0);
printf_filtered ("%-3d ", b->number);
annotate_field (1);
if ((int)b->type > (sizeof(bptypes)/sizeof(bptypes[0]))
|| ((int) b->type != bptypes[(int) b->type].type))
internal_error ("bptypes table does not describe type #%d.",
(int)b->type);
printf_filtered ("%-14s ", bptypes[(int)b->type].description);
annotate_field (2);
printf_filtered ("%-4s ", bpdisps[(int)b->disposition]);
annotate_field (3);
printf_filtered ("%-3c ", bpenables[(int)b->enable]);
...
This is the new version:
annotate_record ();
ui_out_tuple_begin (uiout, "bkpt");
annotate_field (0);
ui_out_field_int (uiout, "number", b->number);
annotate_field (1);
if (((int) b->type > (sizeof (bptypes) / sizeof (bptypes[0])))
|| ((int) b->type != bptypes[(int) b->type].type))
internal_error ("bptypes table does not describe type #%d.",
(int) b->type);
ui_out_field_string (uiout, "type", bptypes[(int)b->type].description);
annotate_field (2);
ui_out_field_string (uiout, "disp", bpdisps[(int)b->disposition]);
annotate_field (3);
ui_out_field_fmt (uiout, "enabled", "%c", bpenables[(int)b->enable]);
...
This example, also from `print_one_breakpoint', shows how to produce
a complicated output field using the `print_expression' functions which
requires a stream to be passed. It also shows how to automate stream
destruction with cleanups. The original code was:
annotate_field (5);
print_expression (b->exp, gdb_stdout);
The new version is:
struct ui_stream *stb = ui_out_stream_new (uiout);
struct cleanup *old_chain = make_cleanup_ui_out_stream_delete (stb);
...
annotate_field (5);
print_expression (b->exp, stb->stream);
ui_out_field_stream (uiout, "what", local_stream);
This example, also from `print_one_breakpoint', shows how to use
`ui_out_text' and `ui_out_field_string'. The original code was:
annotate_field (5);
if (b->dll_pathname == NULL)
printf_filtered ("<any library> ");
else
printf_filtered ("library \"%s\" ", b->dll_pathname);
It became:
annotate_field (5);
if (b->dll_pathname == NULL)
{
ui_out_field_string (uiout, "what", "<any library>");
ui_out_spaces (uiout, 1);
}
else
{
ui_out_text (uiout, "library \"");
ui_out_field_string (uiout, "what", b->dll_pathname);
ui_out_text (uiout, "\" ");
}
The following example from `print_one_breakpoint' shows how to use
`ui_out_field_int' and `ui_out_spaces'. The original code was:
annotate_field (5);
if (b->forked_inferior_pid != 0)
printf_filtered ("process %d ", b->forked_inferior_pid);
It became:
annotate_field (5);
if (b->forked_inferior_pid != 0)
{
ui_out_text (uiout, "process ");
ui_out_field_int (uiout, "what", b->forked_inferior_pid);
ui_out_spaces (uiout, 1);
}
Here's an example of using `ui_out_field_string'. The original code
was:
annotate_field (5);
if (b->exec_pathname != NULL)
printf_filtered ("program \"%s\" ", b->exec_pathname);
It became:
annotate_field (5);
if (b->exec_pathname != NULL)
{
ui_out_text (uiout, "program \"");
ui_out_field_string (uiout, "what", b->exec_pathname);
ui_out_text (uiout, "\" ");
}
Finally, here's an example of printing an address. The original
code:
annotate_field (4);
printf_filtered ("%s ",
hex_string_custom ((unsigned long) b->address, 8));
It became:
annotate_field (4);
ui_out_field_core_addr (uiout, "Address", b->address);
4.3 Console Printing
====================
4.4 TUI
=======
---------- Footnotes ----------
(1) The function cast is not portable ISO C.
(2) As of this writing (April 2001), setting verbosity level is not
yet implemented, and is always returned as zero. So calling
`ui_out_message' with a VERBOSITY argument more than zero will cause
the message to never be printed.
File: gdbint.info, Node: libgdb, Next: Values, Prev: User Interface, Up: Top
5 libgdb
********
5.1 libgdb 1.0
==============
`libgdb' 1.0 was an abortive project of years ago. The theory was to
provide an API to GDB's functionality.
5.2 libgdb 2.0
==============
`libgdb' 2.0 is an ongoing effort to update GDB so that is better able
to support graphical and other environments.
Since `libgdb' development is on-going, its architecture is still
evolving. The following components have so far been identified:
* Observer - `gdb-events.h'.
* Builder - `ui-out.h'
* Event Loop - `event-loop.h'
* Library - `gdb.h'
The model that ties these components together is described below.
5.3 The `libgdb' Model
======================
A client of `libgdb' interacts with the library in two ways.
* As an observer (using `gdb-events') receiving notifications from
`libgdb' of any internal state changes (break point changes, run
state, etc).
* As a client querying `libgdb' (using the `ui-out' builder) to
obtain various status values from GDB.
Since `libgdb' could have multiple clients (e.g., a GUI supporting
the existing GDB CLI), those clients must co-operate when controlling
`libgdb'. In particular, a client must ensure that `libgdb' is idle
(i.e. no other client is using `libgdb') before responding to a
`gdb-event' by making a query.
5.4 CLI support
===============
At present GDB's CLI is very much entangled in with the core of
`libgdb'. Consequently, a client wishing to include the CLI in their
interface needs to carefully co-ordinate its own and the CLI's
requirements.
It is suggested that the client set `libgdb' up to be bi-modal
(alternate between CLI and client query modes). The notes below sketch
out the theory:
* The client registers itself as an observer of `libgdb'.
* The client create and install `cli-out' builder using its own
versions of the `ui-file' `gdb_stderr', `gdb_stdtarg' and
`gdb_stdout' streams.
* The client creates a separate custom `ui-out' builder that is only
used while making direct queries to `libgdb'.
When the client receives input intended for the CLI, it simply
passes it along. Since the `cli-out' builder is installed by default,
all the CLI output in response to that command is routed (pronounced
rooted) through to the client controlled `gdb_stdout' et. al. streams.
At the same time, the client is kept abreast of internal changes by
virtue of being a `libgdb' observer.
The only restriction on the client is that it must wait until
`libgdb' becomes idle before initiating any queries (using the client's
custom builder).
5.5 `libgdb' components
=======================
Observer - `gdb-events.h'
-------------------------
`gdb-events' provides the client with a very raw mechanism that can be
used to implement an observer. At present it only allows for one
observer and that observer must, internally, handle the need to delay
the processing of any event notifications until after `libgdb' has
finished the current command.
Builder - `ui-out.h'
--------------------
`ui-out' provides the infrastructure necessary for a client to create a
builder. That builder is then passed down to `libgdb' when doing any
queries.
Event Loop - `event-loop.h'
---------------------------
`event-loop', currently non-re-entrant, provides a simple event loop.
A client would need to either plug its self into this loop or,
implement a new event-loop that GDB would use.
The event-loop will eventually be made re-entrant. This is so that
GDB can better handle the problem of some commands blocking instead of
returning.
Library - `gdb.h'
-----------------
`libgdb' is the most obvious component of this system. It provides the
query interface. Each function is parameterized by a `ui-out' builder.
The result of the query is constructed using that builder before the
query function returns.
File: gdbint.info, Node: Values, Next: Stack Frames, Prev: libgdb, Up: Top
6 Values
********
6.1 Values
==========
GDB uses `struct value', or "values", as an internal abstraction for
the representation of a variety of inferior objects and GDB convenience
objects.
Values have an associated `struct type', that describes a virtual
view of the raw data or object stored in or accessed through the value.
A value is in addition discriminated by its lvalue-ness, given its
`enum lval_type' enumeration type:
``not_lval''
This value is not an lval. It can't be assigned to.
``lval_memory''
This value represents an object in memory.
``lval_register''
This value represents an object that lives in a register.
``lval_internalvar''
Represents the value of an internal variable.
``lval_internalvar_component''
Represents part of a GDB internal variable. E.g., a structure
field.
``lval_computed''
These are "computed" values. They allow creating specialized value
objects for specific purposes, all abstracted away from the core
value support code. The creator of such a value writes specialized
functions to handle the reading and writing to/from the value's
backend data, and optionally, a "copy operator" and a "destructor".
Pointers to these functions are stored in a `struct lval_funcs'
instance (declared in `value.h'), and passed to the
`allocate_computed_value' function, as in the example below.
static void
nil_value_read (struct value *v)
{
/* This callback reads data from some backend, and stores it in V.
In this case, we always read null data. You'll want to fill in
something more interesting. */
memset (value_contents_all_raw (v),
value_offset (v),
TYPE_LENGTH (value_type (v)));
}
static void
nil_value_write (struct value *v, struct value *fromval)
{
/* Takes the data from FROMVAL and stores it in the backend of V. */
to_oblivion (value_contents_all_raw (fromval),
value_offset (v),
TYPE_LENGTH (value_type (fromval)));
}
static struct lval_funcs nil_value_funcs =
{
nil_value_read,
nil_value_write
};
struct value *
make_nil_value (void)
{
struct type *type;
struct value *v;
type = make_nils_type ();
v = allocate_computed_value (type, &nil_value_funcs, NULL);
return v;
}
See the implementation of the `$_siginfo' convenience variable in
`infrun.c' as a real example use of lval_computed.
File: gdbint.info, Node: Stack Frames, Next: Symbol Handling, Prev: Values, Up: Top
7 Stack Frames
**************
A frame is a construct that GDB uses to keep track of calling and
called functions.
GDB's frame model, a fresh design, was implemented with the need to
support DWARF's Call Frame Information in mind. In fact, the term
"unwind" is taken directly from that specification. Developers wishing
to learn more about unwinders, are encouraged to read the DWARF
specification, available from `http://www.dwarfstd.org'.
GDB's model is that you find a frame's registers by "unwinding" them
from the next younger frame. That is, `get_frame_register' which
returns the value of a register in frame #1 (the next-to-youngest
frame), is implemented by calling frame #0's `frame_register_unwind'
(the youngest frame). But then the obvious question is: how do you
access the registers of the youngest frame itself?
To answer this question, GDB has the "sentinel" frame, the "-1st"
frame. Unwinding registers from the sentinel frame gives you the
current values of the youngest real frame's registers. If F is a
sentinel frame, then `get_frame_type (F) == SENTINEL_FRAME'.
7.1 Selecting an Unwinder
=========================
The architecture registers a list of frame unwinders (`struct
frame_unwind'), using the functions `frame_unwind_prepend_unwinder' and
`frame_unwind_append_unwinder'. Each unwinder includes a sniffer.
Whenever GDB needs to unwind a frame (to fetch the previous frame's
registers or the current frame's ID), it calls registered sniffers in
order to find one which recognizes the frame. The first time a sniffer
returns non-zero, the corresponding unwinder is assigned to the frame.
7.2 Unwinding the Frame ID
==========================
Every frame has an associated ID, of type `struct frame_id'. The ID
includes the stack base and function start address for the frame. The
ID persists through the entire life of the frame, including while other
called frames are running; it is used to locate an appropriate `struct
frame_info' from the cache.
Every time the inferior stops, and at various other times, the frame
cache is flushed. Because of this, parts of GDB which need to keep
track of individual frames cannot use pointers to `struct frame_info'.
A frame ID provides a stable reference to a frame, even when the
unwinder must be run again to generate a new `struct frame_info' for
the same frame.
The frame's unwinder's `this_id' method is called to find the ID.
Note that this is different from register unwinding, where the next
frame's `prev_register' is called to unwind this frame's registers.
Both stack base and function address are required to identify the
frame, because a recursive function has the same function address for
two consecutive frames and a leaf function may have the same stack
address as its caller. On some platforms, a third address is part of
the ID to further disambiguate frames--for instance, on IA-64 the
separate register stack address is included in the ID.
An invalid frame ID (`outer_frame_id') returned from the `this_id'
method means to stop unwinding after this frame.
`null_frame_id' is another invalid frame ID which should be used
when there is no frame. For instance, certain breakpoints are attached
to a specific frame, and that frame is identified through its frame ID
(we use this to implement the "finish" command). Using `null_frame_id'
as the frame ID for a given breakpoint means that the breakpoint is not
specific to any frame. The `this_id' method should never return
`null_frame_id'.
7.3 Unwinding Registers
=======================
Each unwinder includes a `prev_register' method. This method takes a
frame, an associated cache pointer, and a register number. It returns
a `struct value *' describing the requested register, as saved by this
frame. This is the value of the register that is current in this
frame's caller.
The returned value must have the same type as the register. It may
have any lvalue type. In most circumstances one of these routines will
generate the appropriate value:
`frame_unwind_got_optimized'
This register was not saved.
`frame_unwind_got_register'
This register was copied into another register in this frame. This
is also used for unchanged registers; they are "copied" into the
same register.
`frame_unwind_got_memory'
This register was saved in memory.
`frame_unwind_got_constant'
This register was not saved, but the unwinder can compute the
previous value some other way.
`frame_unwind_got_address'
Same as `frame_unwind_got_constant', except that the value is a
target address. This is frequently used for the stack pointer,
which is not explicitly saved but has a known offset from this
frame's stack pointer. For architectures with a flat unified
address space, this is generally the same as
`frame_unwind_got_constant'.
File: gdbint.info, Node: Symbol Handling, Next: Language Support, Prev: Stack Frames, Up: Top
8 Symbol Handling
*****************
Symbols are a key part of GDB's operation. Symbols include variables,
functions, and types.
Symbol information for a large program can be truly massive, and
reading of symbol information is one of the major performance
bottlenecks in GDB; it can take many minutes to process it all.
Studies have shown that nearly all the time spent is computational,
rather than file reading.
One of the ways for GDB to provide a good user experience is to
start up quickly, taking no more than a few seconds. It is simply not
possible to process all of a program's debugging info in that time, and
so we attempt to handle symbols incrementally. For instance, we create
"partial symbol tables" consisting of only selected symbols, and only
expand them to full symbol tables when necessary.
8.1 Symbol Reading
==================
GDB reads symbols from "symbol files". The usual symbol file is the
file containing the program which GDB is debugging. GDB can be
directed to use a different file for symbols (with the `symbol-file'
command), and it can also read more symbols via the `add-file' and
`load' commands. In addition, it may bring in more symbols while
loading shared libraries.
Symbol files are initially opened by code in `symfile.c' using the
BFD library (*note Support Libraries::). BFD identifies the type of
the file by examining its header. `find_sym_fns' then uses this
identification to locate a set of symbol-reading functions.
Symbol-reading modules identify themselves to GDB by calling
`add_symtab_fns' during their module initialization. The argument to
`add_symtab_fns' is a `struct sym_fns' which contains the name (or name
prefix) of the symbol format, the length of the prefix, and pointers to
four functions. These functions are called at various times to process
symbol files whose identification matches the specified prefix.
The functions supplied by each module are:
`XYZ_symfile_init(struct sym_fns *sf)'
Called from `symbol_file_add' when we are about to read a new
symbol file. This function should clean up any internal state
(possibly resulting from half-read previous files, for example)
and prepare to read a new symbol file. Note that the symbol file
which we are reading might be a new "main" symbol file, or might
be a secondary symbol file whose symbols are being added to the
existing symbol table.
The argument to `XYZ_symfile_init' is a newly allocated `struct
sym_fns' whose `bfd' field contains the BFD for the new symbol
file being read. Its `private' field has been zeroed, and can be
modified as desired. Typically, a struct of private information
will be `malloc''d, and a pointer to it will be placed in the
`private' field.
There is no result from `XYZ_symfile_init', but it can call
`error' if it detects an unavoidable problem.
`XYZ_new_init()'
Called from `symbol_file_add' when discarding existing symbols.
This function needs only handle the symbol-reading module's
internal state; the symbol table data structures visible to the
rest of GDB will be discarded by `symbol_file_add'. It has no
arguments and no result. It may be called after
`XYZ_symfile_init', if a new symbol table is being read, or may be
called alone if all symbols are simply being discarded.
`XYZ_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)'
Called from `symbol_file_add' to actually read the symbols from a
symbol-file into a set of psymtabs or symtabs.
`sf' points to the `struct sym_fns' originally passed to
`XYZ_sym_init' for possible initialization. `addr' is the offset
between the file's specified start address and its true address in
memory. `mainline' is 1 if this is the main symbol table being
read, and 0 if a secondary symbol file (e.g., shared library or
dynamically loaded file) is being read.
In addition, if a symbol-reading module creates psymtabs when
XYZ_symfile_read is called, these psymtabs will contain a pointer to a
function `XYZ_psymtab_to_symtab', which can be called from any point in
the GDB symbol-handling code.
`XYZ_psymtab_to_symtab (struct partial_symtab *pst)'
Called from `psymtab_to_symtab' (or the `PSYMTAB_TO_SYMTAB' macro)
if the psymtab has not already been read in and had its
`pst->symtab' pointer set. The argument is the psymtab to be
fleshed-out into a symtab. Upon return, `pst->readin' should have
been set to 1, and `pst->symtab' should contain a pointer to the
new corresponding symtab, or zero if there were no symbols in that
part of the symbol file.
8.2 Partial Symbol Tables
=========================
GDB has three types of symbol tables:
* Full symbol tables ("symtabs"). These contain the main
information about symbols and addresses.
* Partial symbol tables ("psymtabs"). These contain enough
information to know when to read the corresponding part of the full
symbol table.
* Minimal symbol tables ("msymtabs"). These contain information
gleaned from non-debugging symbols.
This section describes partial symbol tables.
A psymtab is constructed by doing a very quick pass over an
executable file's debugging information. Small amounts of information
are extracted--enough to identify which parts of the symbol table will
need to be re-read and fully digested later, when the user needs the
information. The speed of this pass causes GDB to start up very
quickly. Later, as the detailed rereading occurs, it occurs in small
pieces, at various times, and the delay therefrom is mostly invisible to
the user.
The symbols that show up in a file's psymtab should be, roughly,
those visible to the debugger's user when the program is not running
code from that file. These include external symbols and types, static
symbols and types, and `enum' values declared at file scope.
The psymtab also contains the range of instruction addresses that the
full symbol table would represent.
The idea is that there are only two ways for the user (or much of the
code in the debugger) to reference a symbol:
* By its address (e.g., execution stops at some address which is
inside a function in this file). The address will be noticed to
be in the range of this psymtab, and the full symtab will be read
in. `find_pc_function', `find_pc_line', and other `find_pc_...'
functions handle this.
* By its name (e.g., the user asks to print a variable, or set a
breakpoint on a function). Global names and file-scope names will
be found in the psymtab, which will cause the symtab to be pulled
in. Local names will have to be qualified by a global name, or a
file-scope name, in which case we will have already read in the
symtab as we evaluated the qualifier. Or, a local symbol can be
referenced when we are "in" a local scope, in which case the first
case applies. `lookup_symbol' does most of the work here.
The only reason that psymtabs exist is to cause a symtab to be read
in at the right moment. Any symbol that can be elided from a psymtab,
while still causing that to happen, should not appear in it. Since
psymtabs don't have the idea of scope, you can't put local symbols in
them anyway. Psymtabs don't have the idea of the type of a symbol,
either, so types need not appear, unless they will be referenced by
name.
It is a bug for GDB to behave one way when only a psymtab has been
read, and another way if the corresponding symtab has been read in.
Such bugs are typically caused by a psymtab that does not contain all
the visible symbols, or which has the wrong instruction address ranges.
The psymtab for a particular section of a symbol file (objfile)
could be thrown away after the symtab has been read in. The symtab
should always be searched before the psymtab, so the psymtab will never
be used (in a bug-free environment). Currently, psymtabs are allocated
on an obstack, and all the psymbols themselves are allocated in a pair
of large arrays on an obstack, so there is little to be gained by
trying to free them unless you want to do a lot more work.
Whether or not psymtabs are created depends on the objfile's symbol
reader. The core of GDB hides the details of partial symbols and
partial symbol tables behind a set of function pointers known as the
"quick symbol functions". These are documented in `symfile.h'.
8.3 Types
=========
Fundamental Types (e.g., `FT_VOID', `FT_BOOLEAN').
--------------------------------------------------
These are the fundamental types that GDB uses internally. Fundamental
types from the various debugging formats (stabs, ELF, etc) are mapped
into one of these. They are basically a union of all fundamental types
that GDB knows about for all the languages that GDB knows about.
Type Codes (e.g., `TYPE_CODE_PTR', `TYPE_CODE_ARRAY').
------------------------------------------------------
Each time GDB builds an internal type, it marks it with one of these
types. The type may be a fundamental type, such as `TYPE_CODE_INT', or
a derived type, such as `TYPE_CODE_PTR' which is a pointer to another
type. Typically, several `FT_*' types map to one `TYPE_CODE_*' type,
and are distinguished by other members of the type struct, such as
whether the type is signed or unsigned, and how many bits it uses.
Builtin Types (e.g., `builtin_type_void', `builtin_type_char').
---------------------------------------------------------------
These are instances of type structs that roughly correspond to
fundamental types and are created as global types for GDB to use for
various ugly historical reasons. We eventually want to eliminate
these. Note for example that `builtin_type_int' initialized in
`gdbtypes.c' is basically the same as a `TYPE_CODE_INT' type that is
initialized in `c-lang.c' for an `FT_INTEGER' fundamental type. The
difference is that the `builtin_type' is not associated with any
particular objfile, and only one instance exists, while `c-lang.c'
builds as many `TYPE_CODE_INT' types as needed, with each one
associated with some particular objfile.
8.4 Object File Formats
=======================
8.4.1 a.out
-----------
The `a.out' format is the original file format for Unix. It consists
of three sections: `text', `data', and `bss', which are for program
code, initialized data, and uninitialized data, respectively.
The `a.out' format is so simple that it doesn't have any reserved
place for debugging information. (Hey, the original Unix hackers used
`adb', which is a machine-language debugger!) The only debugging
format for `a.out' is stabs, which is encoded as a set of normal
symbols with distinctive attributes.
The basic `a.out' reader is in `dbxread.c'.
8.4.2 COFF
----------
The COFF format was introduced with System V Release 3 (SVR3) Unix.
COFF files may have multiple sections, each prefixed by a header. The
number of sections is limited.
The COFF specification includes support for debugging. Although this
was a step forward, the debugging information was woefully limited.
For instance, it was not possible to represent code that came from an
included file. GNU's COFF-using configs often use stabs-type info,
encapsulated in special sections.
The COFF reader is in `coffread.c'.
8.4.3 ECOFF
-----------
ECOFF is an extended COFF originally introduced for Mips and Alpha
workstations.
The basic ECOFF reader is in `mipsread.c'.
8.4.4 XCOFF
-----------
The IBM RS/6000 running AIX uses an object file format called XCOFF.
The COFF sections, symbols, and line numbers are used, but debugging
symbols are `dbx'-style stabs whose strings are located in the `.debug'
section (rather than the string table). For more information, see
*note Top: (stabs)Top.
The shared library scheme has a clean interface for figuring out what
shared libraries are in use, but the catch is that everything which
refers to addresses (symbol tables and breakpoints at least) needs to be
relocated for both shared libraries and the main executable. At least
using the standard mechanism this can only be done once the program has
been run (or the core file has been read).
8.4.5 PE
--------
Windows 95 and NT use the PE ("Portable Executable") format for their
executables. PE is basically COFF with additional headers.
While BFD includes special PE support, GDB needs only the basic COFF
reader.
8.4.6 ELF
---------
The ELF format came with System V Release 4 (SVR4) Unix. ELF is
similar to COFF in being organized into a number of sections, but it
removes many of COFF's limitations. Debugging info may be either stabs
encapsulated in ELF sections, or more commonly these days, DWARF.
The basic ELF reader is in `elfread.c'.
8.4.7 SOM
---------
SOM is HP's object file and debug format (not to be confused with IBM's
SOM, which is a cross-language ABI).
The SOM reader is in `somread.c'.
8.5 Debugging File Formats
==========================
This section describes characteristics of debugging information that
are independent of the object file format.
8.5.1 stabs
-----------
`stabs' started out as special symbols within the `a.out' format.
Since then, it has been encapsulated into other file formats, such as
COFF and ELF.
While `dbxread.c' does some of the basic stab processing, including
for encapsulated versions, `stabsread.c' does the real work.
8.5.2 COFF
----------
The basic COFF definition includes debugging information. The level of
support is minimal and non-extensible, and is not often used.
8.5.3 Mips debug (Third Eye)
----------------------------
ECOFF includes a definition of a special debug format.
The file `mdebugread.c' implements reading for this format.
8.5.4 DWARF 2
-------------
DWARF 2 is an improved but incompatible version of DWARF 1.
The DWARF 2 reader is in `dwarf2read.c'.
8.5.5 Compressed DWARF 2
------------------------
Compressed DWARF 2 is not technically a separate debugging format, but
merely DWARF 2 debug information that has been compressed. In this
format, every object-file section holding DWARF 2 debugging information
is compressed and prepended with a header. (The section is also
typically renamed, so a section called `.debug_info' in a DWARF 2
binary would be called `.zdebug_info' in a compressed DWARF 2 binary.)
The header is 12 bytes long:
* 4 bytes: the literal string "ZLIB"
* 8 bytes: the uncompressed size of the section, in big-endian byte
order.
The same reader is used for both compressed an normal DWARF 2 info.
Section decompression is done in `zlib_decompress_section' in
`dwarf2read.c'.
8.5.6 DWARF 3
-------------
DWARF 3 is an improved version of DWARF 2.
8.5.7 SOM
---------
Like COFF, the SOM definition includes debugging information.
8.6 Adding a New Symbol Reader to GDB
=====================================
If you are using an existing object file format (`a.out', COFF, ELF,
etc), there is probably little to be done.
If you need to add a new object file format, you must first add it to
BFD. This is beyond the scope of this document.
You must then arrange for the BFD code to provide access to the
debugging symbols. Generally GDB will have to call swapping routines
from BFD and a few other BFD internal routines to locate the debugging
information. As much as possible, GDB should not depend on the BFD
internal data structures.
For some targets (e.g., COFF), there is a special transfer vector
used to call swapping routines, since the external data structures on
various platforms have different sizes and layouts. Specialized
routines that will only ever be implemented by one object file format
may be called directly. This interface should be described in a file
`bfd/libXYZ.h', which is included by GDB.
8.7 Memory Management for Symbol Files
======================================
Most memory associated with a loaded symbol file is stored on its
`objfile_obstack'. This includes symbols, types, namespace data, and
other information produced by the symbol readers.
Because this data lives on the objfile's obstack, it is automatically
released when the objfile is unloaded or reloaded. Therefore one
objfile must not reference symbol or type data from another objfile;
they could be unloaded at different times.
User convenience variables, et cetera, have associated types.
Normally these types live in the associated objfile. However, when the
objfile is unloaded, those types are deep copied to global memory, so
that the values of the user variables and history items are not lost.
File: gdbint.info, Node: Language Support, Next: Host Definition, Prev: Symbol Handling, Up: Top
9 Language Support
******************
GDB's language support is mainly driven by the symbol reader, although
it is possible for the user to set the source language manually.
GDB chooses the source language by looking at the extension of the
file recorded in the debug info; `.c' means C, `.f' means Fortran, etc.
It may also use a special-purpose language identifier if the debug
format supports it, like with DWARF.
9.1 Adding a Source Language to GDB
===================================
To add other languages to GDB's expression parser, follow the following
steps:
_Create the expression parser._
This should reside in a file `LANG-exp.y'. Routines for building
parsed expressions into a `union exp_element' list are in
`parse.c'.
Since we can't depend upon everyone having Bison, and YACC produces
parsers that define a bunch of global names, the following lines
*must* be included at the top of the YACC parser, to prevent the
various parsers from defining the same global names:
#define yyparse LANG_parse
#define yylex LANG_lex
#define yyerror LANG_error
#define yylval LANG_lval
#define yychar LANG_char
#define yydebug LANG_debug
#define yypact LANG_pact
#define yyr1 LANG_r1
#define yyr2 LANG_r2
#define yydef LANG_def
#define yychk LANG_chk
#define yypgo LANG_pgo
#define yyact LANG_act
#define yyexca LANG_exca
#define yyerrflag LANG_errflag
#define yynerrs LANG_nerrs
At the bottom of your parser, define a `struct language_defn' and
initialize it with the right values for your language. Define an
`initialize_LANG' routine and have it call
`add_language(LANG_language_defn)' to tell the rest of GDB that
your language exists. You'll need some other supporting variables
and functions, which will be used via pointers from your
`LANG_language_defn'. See the declaration of `struct
language_defn' in `language.h', and the other `*-exp.y' files, for
more information.
_Add any evaluation routines, if necessary_
If you need new opcodes (that represent the operations of the
language), add them to the enumerated type in `expression.h'. Add
support code for these operations in the `evaluate_subexp' function
defined in the file `eval.c'. Add cases for new opcodes in two
functions from `parse.c': `prefixify_subexp' and
`length_of_subexp'. These compute the number of `exp_element's
that a given operation takes up.
_Update some existing code_
Add an enumerated identifier for your language to the enumerated
type `enum language' in `defs.h'.
Update the routines in `language.c' so your language is included.
These routines include type predicates and such, which (in some
cases) are language dependent. If your language does not appear
in the switch statement, an error is reported.
Also included in `language.c' is the code that updates the variable
`current_language', and the routines that translate the
`language_LANG' enumerated identifier into a printable string.
Update the function `_initialize_language' to include your
language. This function picks the default language upon startup,
so is dependent upon which languages that GDB is built for.
Update `allocate_symtab' in `symfile.c' and/or symbol-reading code
so that the language of each symtab (source file) is set properly.
This is used to determine the language to use at each stack frame
level. Currently, the language is set based upon the extension of
the source file. If the language can be better inferred from the
symbol information, please set the language of the symtab in the
symbol-reading code.
Add helper code to `print_subexp' (in `expprint.c') to handle any
new expression opcodes you have added to `expression.h'. Also,
add the printed representations of your operators to
`op_print_tab'.
_Add a place of call_
Add a call to `LANG_parse()' and `LANG_error' in `parse_exp_1'
(defined in `parse.c').
_Edit `Makefile.in'_
Add dependencies in `Makefile.in'. Make sure you update the macro
variables such as `HFILES' and `OBJS', otherwise your code may not
get linked in, or, worse yet, it may not get `tar'red into the
distribution!
File: gdbint.info, Node: Host Definition, Next: Target Architecture Definition, Prev: Language Support, Up: Top
10 Host Definition
******************
With the advent of Autoconf, it's rarely necessary to have host
definition machinery anymore. The following information is provided,
mainly, as an historical reference.
10.1 Adding a New Host
======================
GDB's host configuration support normally happens via Autoconf. New
host-specific definitions should not be needed. Older hosts GDB still
use the host-specific definitions and files listed below, but these
mostly exist for historical reasons, and will eventually disappear.
`gdb/config/ARCH/XYZ.mh'
This file is a Makefile fragment that once contained both host and
native configuration information (*note Native Debugging::) for the
machine XYZ. The host configuration information is now handled by
Autoconf.
Host configuration information included definitions for `CC',
`SYSV_DEFINE', `XM_CFLAGS', `XM_ADD_FILES', `XM_CLIBS',
`XM_CDEPS', etc.; see `Makefile.in'.
New host-only configurations do not need this file.
(Files named `gdb/config/ARCH/xm-XYZ.h' were once used to define
host-specific macros, but were no longer needed and have all been
removed.)
Generic Host Support Files
--------------------------
There are some "generic" versions of routines that can be used by
various systems.
`ser-unix.c'
This contains serial line support for Unix systems. It is
included by default on all Unix-like hosts.
`ser-pipe.c'
This contains serial pipe support for Unix systems. It is
included by default on all Unix-like hosts.
`ser-mingw.c'
This contains serial line support for 32-bit programs running under
Windows using MinGW.
`ser-go32.c'
This contains serial line support for 32-bit programs running
under DOS, using the DJGPP (a.k.a. GO32) execution environment.
`ser-tcp.c'
This contains generic TCP support using sockets. It is included by
default on all Unix-like hosts and with MinGW.
10.2 Host Conditionals
======================
When GDB is configured and compiled, various macros are defined or left
undefined, to control compilation based on the attributes of the host
system. While formerly they could be set in host-specific header
files, at present they can be changed only by setting `CFLAGS' when
building, or by editing the source code.
These macros and their meanings (or if the meaning is not documented
here, then one of the source files where they are used is indicated)
are:
`GDBINIT_FILENAME'
The default name of GDB's initialization file (normally
`.gdbinit').
`SIGWINCH_HANDLER'
If your host defines `SIGWINCH', you can define this to be the name
of a function to be called if `SIGWINCH' is received.
`SIGWINCH_HANDLER_BODY'
Define this to expand into code that will define the function
named by the expansion of `SIGWINCH_HANDLER'.
`CRLF_SOURCE_FILES'
Define this if host files use `\r\n' rather than `\n' as a line
terminator. This will cause source file listings to omit `\r'
characters when printing and it will allow `\r\n' line endings of
files which are "sourced" by gdb. It must be possible to open
files in binary mode using `O_BINARY' or, for fopen, `"rb"'.
`DEFAULT_PROMPT'
The default value of the prompt string (normally `"(gdb) "').
`DEV_TTY'
The name of the generic TTY device, defaults to `"/dev/tty"'.
`ISATTY'
Substitute for isatty, if not available.
`FOPEN_RB'
Define this if binary files are opened the same way as text files.
`CC_HAS_LONG_LONG'
Define this if the host C compiler supports `long long'. This is
set by the `configure' script.
`PRINTF_HAS_LONG_LONG'
Define this if the host can handle printing of long long integers
via the printf format conversion specifier `ll'. This is set by
the `configure' script.
`LSEEK_NOT_LINEAR'
Define this if `lseek (n)' does not necessarily move to byte number
`n' in the file. This is only used when reading source files. It
is normally faster to define `CRLF_SOURCE_FILES' when possible.
`lint'
Define this to help placate `lint' in some situations.
`volatile'
Define this to override the defaults of `__volatile__' or `/**/'.
File: gdbint.info, Node: Target Architecture Definition, Next: Target Descriptions, Prev: Host Definition, Up: Top
11 Target Architecture Definition
*********************************
GDB's target architecture defines what sort of machine-language
programs GDB can work with, and how it works with them.
The target architecture object is implemented as the C structure
`struct gdbarch *'. The structure, and its methods, are generated
using the Bourne shell script `gdbarch.sh'.
* Menu:
* OS ABI Variant Handling::
* Initialize New Architecture::
* Registers and Memory::
* Pointers and Addresses::
* Address Classes::
* Register Representation::
* Frame Interpretation::
* Inferior Call Setup::
* Adding support for debugging core files::
* Defining Other Architecture Features::
* Adding a New Target::
File: gdbint.info, Node: OS ABI Variant Handling, Next: Initialize New Architecture, Up: Target Architecture Definition
11.1 Operating System ABI Variant Handling
==========================================
GDB provides a mechanism for handling variations in OS ABIs. An OS ABI
variant may have influence over any number of variables in the target
architecture definition. There are two major components in the OS ABI
mechanism: sniffers and handlers.
A "sniffer" examines a file matching a BFD architecture/flavour pair
(the architecture may be wildcarded) in an attempt to determine the OS
ABI of that file. Sniffers with a wildcarded architecture are
considered to be "generic", while sniffers for a specific architecture
are considered to be "specific". A match from a specific sniffer
overrides a match from a generic sniffer. Multiple sniffers for an
architecture/flavour may exist, in order to differentiate between two
different operating systems which use the same basic file format. The
OS ABI framework provides a generic sniffer for ELF-format files which
examines the `EI_OSABI' field of the ELF header, as well as note
sections known to be used by several operating systems.
A "handler" is used to fine-tune the `gdbarch' structure for the
selected OS ABI. There may be only one handler for a given OS ABI for
each BFD architecture.
The following OS ABI variants are defined in `defs.h':
`GDB_OSABI_UNINITIALIZED'
Used for struct gdbarch_info if ABI is still uninitialized.
`GDB_OSABI_UNKNOWN'
The ABI of the inferior is unknown. The default `gdbarch'
settings for the architecture will be used.
`GDB_OSABI_SVR4'
UNIX System V Release 4.
`GDB_OSABI_HURD'
GNU using the Hurd kernel.
`GDB_OSABI_SOLARIS'
Sun Solaris.
`GDB_OSABI_OSF1'
OSF/1, including Digital UNIX and Compaq Tru64 UNIX.
`GDB_OSABI_LINUX'
GNU using the Linux kernel.
`GDB_OSABI_FREEBSD_AOUT'
FreeBSD using the `a.out' executable format.
`GDB_OSABI_FREEBSD_ELF'
FreeBSD using the ELF executable format.
`GDB_OSABI_NETBSD_AOUT'
NetBSD using the `a.out' executable format.
`GDB_OSABI_NETBSD_ELF'
NetBSD using the ELF executable format.
`GDB_OSABI_OPENBSD_ELF'
OpenBSD using the ELF executable format.
`GDB_OSABI_WINCE'
Windows CE.
`GDB_OSABI_GO32'
DJGPP.
`GDB_OSABI_IRIX'
Irix.
`GDB_OSABI_INTERIX'
Interix (Posix layer for MS-Windows systems).
`GDB_OSABI_HPUX_ELF'
HP/UX using the ELF executable format.
`GDB_OSABI_HPUX_SOM'
HP/UX using the SOM executable format.
`GDB_OSABI_QNXNTO'
QNX Neutrino.
`GDB_OSABI_CYGWIN'
Cygwin.
`GDB_OSABI_AIX'
AIX.
Here are the functions that make up the OS ABI framework:
-- Function: const char * gdbarch_osabi_name (enum gdb_osabi OSABI)
Return the name of the OS ABI corresponding to OSABI.
-- Function: void gdbarch_register_osabi (enum bfd_architecture ARCH,
unsigned long MACHINE, enum gdb_osabi OSABI, void
(*INIT_OSABI)(struct gdbarch_info INFO, struct gdbarch
*GDBARCH))
Register the OS ABI handler specified by INIT_OSABI for the
architecture, machine type and OS ABI specified by ARCH, MACHINE
and OSABI. In most cases, a value of zero for the machine type,
which implies the architecture's default machine type, will
suffice.
-- Function: void gdbarch_register_osabi_sniffer (enum
bfd_architecture ARCH, enum bfd_flavour FLAVOUR, enum
gdb_osabi (*SNIFFER)(bfd *ABFD))
Register the OS ABI file sniffer specified by SNIFFER for the BFD
architecture/flavour pair specified by ARCH and FLAVOUR. If ARCH
is `bfd_arch_unknown', the sniffer is considered to be generic,
and is allowed to examine FLAVOUR-flavoured files for any
architecture.
-- Function: enum gdb_osabi gdbarch_lookup_osabi (bfd *ABFD)
Examine the file described by ABFD to determine its OS ABI. The
value `GDB_OSABI_UNKNOWN' is returned if the OS ABI cannot be
determined.
-- Function: void gdbarch_init_osabi (struct gdbarch info INFO, struct
gdbarch *GDBARCH, enum gdb_osabi OSABI)
Invoke the OS ABI handler corresponding to OSABI to fine-tune the
`gdbarch' structure specified by GDBARCH. If a handler
corresponding to OSABI has not been registered for GDBARCH's
architecture, a warning will be issued and the debugging session
will continue with the defaults already established for GDBARCH.
-- Function: void generic_elf_osabi_sniff_abi_tag_sections (bfd *ABFD,
asection *SECT, void *OBJ)
Helper routine for ELF file sniffers. Examine the file described
by ABFD and look at ABI tag note sections to determine the OS ABI
from the note. This function should be called via
`bfd_map_over_sections'.
File: gdbint.info, Node: Initialize New Architecture, Next: Registers and Memory, Prev: OS ABI Variant Handling, Up: Target Architecture Definition
11.2 Initializing a New Architecture
====================================
* Menu:
* How an Architecture is Represented::
* Looking Up an Existing Architecture::
* Creating a New Architecture::
File: gdbint.info, Node: How an Architecture is Represented, Next: Looking Up an Existing Architecture, Up: Initialize New Architecture
11.2.1 How an Architecture is Represented
-----------------------------------------
Each `gdbarch' is associated with a single BFD architecture, via a
`bfd_arch_ARCH' in the `bfd_architecture' enumeration. The `gdbarch'
is registered by a call to `register_gdbarch_init', usually from the
file's `_initialize_FILENAME' routine, which will be automatically
called during GDB startup. The arguments are a BFD architecture
constant and an initialization function.
A GDB description for a new architecture, ARCH is created by
defining a global function `_initialize_ARCH_tdep', by convention in
the source file `ARCH-tdep.c'. For example, in the case of the
OpenRISC 1000, this function is called `_initialize_or1k_tdep' and is
found in the file `or1k-tdep.c'.
The resulting object files containing the implementation of the
`_initialize_ARCH_tdep' function are specified in the GDB
`configure.tgt' file, which includes a large case statement pattern
matching against the `--target' option of the `configure' script. The
new `struct gdbarch' is created within the `_initialize_ARCH_tdep'
function by calling `gdbarch_register':
void gdbarch_register (enum bfd_architecture ARCHITECTURE,
gdbarch_init_ftype *INIT_FUNC,
gdbarch_dump_tdep_ftype *TDEP_DUMP_FUNC);
The ARCHITECTURE will identify the unique BFD to be associated with
this `gdbarch'. The INIT_FUNC funciton is called to create and return
the new `struct gdbarch'. The TDEP_DUMP_FUNC function will dump the
target specific details associated with this architecture.
For example the function `_initialize_or1k_tdep' creates its
architecture for 32-bit OpenRISC 1000 architectures by calling:
gdbarch_register (bfd_arch_or32, or1k_gdbarch_init, or1k_dump_tdep);
File: gdbint.info, Node: Looking Up an Existing Architecture, Next: Creating a New Architecture, Prev: How an Architecture is Represented, Up: Initialize New Architecture
11.2.2 Looking Up an Existing Architecture
------------------------------------------
The initialization function has this prototype:
static struct gdbarch *
ARCH_gdbarch_init (struct gdbarch_info INFO,
struct gdbarch_list *ARCHES)
The INFO argument contains parameters used to select the correct
architecture, and ARCHES is a list of architectures which have already
been created with the same `bfd_arch_ARCH' value.
The initialization function should first make sure that INFO is
acceptable, and return `NULL' if it is not. Then, it should search
through ARCHES for an exact match to INFO, and return one if found.
Lastly, if no exact match was found, it should create a new
architecture based on INFO and return it.
The lookup is done using `gdbarch_list_lookup_by_info'. It is
passed the list of existing architectures, ARCHES, and the `struct
gdbarch_info', INFO, and returns the first matching architecture it
finds, or `NULL' if none are found. If an architecture is found it can
be returned as the result from the initialization function, otherwise a
new `struct gdbach' will need to be created.
The struct gdbarch_info has the following components:
struct gdbarch_info
{
const struct bfd_arch_info *bfd_arch_info;
int byte_order;
bfd *abfd;
struct gdbarch_tdep_info *tdep_info;
enum gdb_osabi osabi;
const struct target_desc *target_desc;
};
The `bfd_arch_info' member holds the key details about the
architecture. The `byte_order' member is a value in an enumeration
indicating the endianism. The `abfd' member is a pointer to the full
BFD, the `tdep_info' member is additional custom target specific
information, `osabi' identifies which (if any) of a number of operating
specific ABIs are used by this architecture and the `target_desc'
member is a set of name-value pairs with information about register
usage in this target.
When the `struct gdbarch' initialization function is called, not all
the fields are provided--only those which can be deduced from the BFD.
The `struct gdbarch_info', INFO is used as a look-up key with the list
of existing architectures, ARCHES to see if a suitable architecture
already exists. The TDEP_INFO, OSABI and TARGET_DESC fields may be
added before this lookup to refine the search.
Only information in INFO should be used to choose the new
architecture. Historically, INFO could be sparse, and defaults would
be collected from the first element on ARCHES. However, GDB now fills
in INFO more thoroughly, so new `gdbarch' initialization functions
should not take defaults from ARCHES.
File: gdbint.info, Node: Creating a New Architecture, Prev: Looking Up an Existing Architecture, Up: Initialize New Architecture
11.2.3 Creating a New Architecture
----------------------------------
If no architecture is found, then a new architecture must be created,
by calling `gdbarch_alloc' using the supplied `struct gdbarch_info' and
any additional custom target specific information in a `struct
gdbarch_tdep'. The prototype for `gdbarch_alloc' is:
struct gdbarch *gdbarch_alloc (const struct gdbarch_info *INFO,
struct gdbarch_tdep *TDEP);
The newly created struct gdbarch must then be populated. Although
there are default values, in most cases they are not what is required.
For each element, X, there is are a pair of corresponding accessor
functions, one to set the value of that element, `set_gdbarch_X', the
second to either get the value of an element (if it is a variable) or
to apply the element (if it is a function), `gdbarch_X'. Note that
both accessor functions take a pointer to the `struct gdbarch' as first
argument. Populating the new `gdbarch' should use the `set_gdbarch'
functions.
The following sections identify the main elements that should be set
in this way. This is not the complete list, but represents the
functions and elements that must commonly be specified for a new
architecture. Many of the functions and variables are described in the
header file `gdbarch.h'.
This is the main work in defining a new architecture. Implementing
the set of functions to populate the `struct gdbarch'.
`struct gdbarch_tdep' is not defined within GDB--it is up to the
user to define this struct if it is needed to hold custom target
information that is not covered by the standard `struct gdbarch'. For
example with the OpenRISC 1000 architecture it is used to hold the
number of matchpoints available in the target (along with other
information).
If there is no additional target specific information, it can be set
to `NULL'.
File: gdbint.info, Node: Registers and Memory, Next: Pointers and Addresses, Prev: Initialize New Architecture, Up: Target Architecture Definition
11.3 Registers and Memory
=========================
GDB's model of the target machine is rather simple. GDB assumes the
machine includes a bank of registers and a block of memory. Each
register may have a different size.
GDB does not have a magical way to match up with the compiler's idea
of which registers are which; however, it is critical that they do
match up accurately. The only way to make this work is to get accurate
information about the order that the compiler uses, and to reflect that
in the `gdbarch_register_name' and related functions.
GDB can handle big-endian, little-endian, and bi-endian
architectures.
File: gdbint.info, Node: Pointers and Addresses, Next: Address Classes, Prev: Registers and Memory, Up: Target Architecture Definition
11.4 Pointers Are Not Always Addresses
======================================
On almost all 32-bit architectures, the representation of a pointer is
indistinguishable from the representation of some fixed-length number
whose value is the byte address of the object pointed to. On such
machines, the words "pointer" and "address" can be used interchangeably.
However, architectures with smaller word sizes are often cramped for
address space, so they may choose a pointer representation that breaks
this identity, and allows a larger code address space.
For example, the Renesas D10V is a 16-bit VLIW processor whose
instructions are 32 bits long(1). If the D10V used ordinary byte
addresses to refer to code locations, then the processor would only be
able to address 64kb of instructions. However, since instructions must
be aligned on four-byte boundaries, the low two bits of any valid
instruction's byte address are always zero--byte addresses waste two
bits. So instead of byte addresses, the D10V uses word addresses--byte
addresses shifted right two bits--to refer to code. Thus, the D10V can
use 16-bit words to address 256kb of code space.
However, this means that code pointers and data pointers have
different forms on the D10V. The 16-bit word `0xC020' refers to byte
address `0xC020' when used as a data address, but refers to byte address
`0x30080' when used as a code address.
(The D10V also uses separate code and data address spaces, which also
affects the correspondence between pointers and addresses, but we're
going to ignore that here; this example is already too long.)
To cope with architectures like this--the D10V is not the only
one!--GDB tries to distinguish between "addresses", which are byte
numbers, and "pointers", which are the target's representation of an
address of a particular type of data. In the example above, `0xC020'
is the pointer, which refers to one of the addresses `0xC020' or
`0x30080', depending on the type imposed upon it. GDB provides
functions for turning a pointer into an address and vice versa, in the
appropriate way for the current architecture.
Unfortunately, since addresses and pointers are identical on almost
all processors, this distinction tends to bit-rot pretty quickly. Thus,
each time you port GDB to an architecture which does distinguish
between pointers and addresses, you'll probably need to clean up some
architecture-independent code.
Here are functions which convert between pointers and addresses:
-- Function: CORE_ADDR extract_typed_address (void *BUF, struct type
*TYPE)
Treat the bytes at BUF as a pointer or reference of type TYPE, and
return the address it represents, in a manner appropriate for the
current architecture. This yields an address GDB can use to read
target memory, disassemble, etc. Note that BUF refers to a buffer
in GDB's memory, not the inferior's.
For example, if the current architecture is the Intel x86, this
function extracts a little-endian integer of the appropriate
length from BUF and returns it. However, if the current
architecture is the D10V, this function will return a 16-bit
integer extracted from BUF, multiplied by four if TYPE is a
pointer to a function.
If TYPE is not a pointer or reference type, then this function
will signal an internal error.
-- Function: CORE_ADDR store_typed_address (void *BUF, struct type
*TYPE, CORE_ADDR ADDR)
Store the address ADDR in BUF, in the proper format for a pointer
of type TYPE in the current architecture. Note that BUF refers to
a buffer in GDB's memory, not the inferior's.
For example, if the current architecture is the Intel x86, this
function stores ADDR unmodified as a little-endian integer of the
appropriate length in BUF. However, if the current architecture
is the D10V, this function divides ADDR by four if TYPE is a
pointer to a function, and then stores it in BUF.
If TYPE is not a pointer or reference type, then this function
will signal an internal error.
-- Function: CORE_ADDR value_as_address (struct value *VAL)
Assuming that VAL is a pointer, return the address it represents,
as appropriate for the current architecture.
This function actually works on integral values, as well as
pointers. For pointers, it performs architecture-specific
conversions as described above for `extract_typed_address'.
-- Function: CORE_ADDR value_from_pointer (struct type *TYPE,
CORE_ADDR ADDR)
Create and return a value representing a pointer of type TYPE to
the address ADDR, as appropriate for the current architecture.
This function performs architecture-specific conversions as
described above for `store_typed_address'.
Here are two functions which architectures can define to indicate the
relationship between pointers and addresses. These have default
definitions, appropriate for architectures on which all pointers are
simple unsigned byte addresses.
-- Function: CORE_ADDR gdbarch_pointer_to_address (struct gdbarch
*GDBARCH, struct type *TYPE, char *BUF)
Assume that BUF holds a pointer of type TYPE, in the appropriate
format for the current architecture. Return the byte address the
pointer refers to.
This function may safely assume that TYPE is either a pointer or a
C++ reference type.
-- Function: void gdbarch_address_to_pointer (struct gdbarch *GDBARCH,
struct type *TYPE, char *BUF, CORE_ADDR ADDR)
Store in BUF a pointer of type TYPE representing the address ADDR,
in the appropriate format for the current architecture.
This function may safely assume that TYPE is either a pointer or a
C++ reference type.
---------- Footnotes ----------
(1) Some D10V instructions are actually pairs of 16-bit
sub-instructions. However, since you can't jump into the middle of
such a pair, code addresses can only refer to full 32 bit instructions,
which is what matters in this explanation.
File: gdbint.info, Node: Address Classes, Next: Register Representation, Prev: Pointers and Addresses, Up: Target Architecture Definition
11.5 Address Classes
====================
Sometimes information about different kinds of addresses is available
via the debug information. For example, some programming environments
define addresses of several different sizes. If the debug information
distinguishes these kinds of address classes through either the size
info (e.g, `DW_AT_byte_size' in DWARF 2) or through an explicit address
class attribute (e.g, `DW_AT_address_class' in DWARF 2), the following
macros should be defined in order to disambiguate these types within
GDB as well as provide the added information to a GDB user when
printing type expressions.
-- Function: int gdbarch_address_class_type_flags (struct gdbarch
*GDBARCH, int BYTE_SIZE, int DWARF2_ADDR_CLASS)
Returns the type flags needed to construct a pointer type whose
size is BYTE_SIZE and whose address class is DWARF2_ADDR_CLASS.
This function is normally called from within a symbol reader. See
`dwarf2read.c'.
-- Function: char * gdbarch_address_class_type_flags_to_name (struct
gdbarch *GDBARCH, int TYPE_FLAGS)
Given the type flags representing an address class qualifier,
return its name.
-- Function: int gdbarch_address_class_name_to_type_flags (struct
gdbarch *GDBARCH, int NAME, int *TYPE_FLAGS_PTR)
Given an address qualifier name, set the `int' referenced by
TYPE_FLAGS_PTR to the type flags for that address class qualifier.
Since the need for address classes is rather rare, none of the
address class functions are defined by default. Predicate functions
are provided to detect when they are defined.
Consider a hypothetical architecture in which addresses are normally
32-bits wide, but 16-bit addresses are also supported. Furthermore,
suppose that the DWARF 2 information for this architecture simply uses
a `DW_AT_byte_size' value of 2 to indicate the use of one of these
"short" pointers. The following functions could be defined to
implement the address class functions:
somearch_address_class_type_flags (int byte_size,
int dwarf2_addr_class)
{
if (byte_size == 2)
return TYPE_FLAG_ADDRESS_CLASS_1;
else
return 0;
}
static char *
somearch_address_class_type_flags_to_name (int type_flags)
{
if (type_flags & TYPE_FLAG_ADDRESS_CLASS_1)
return "short";
else
return NULL;
}
int
somearch_address_class_name_to_type_flags (char *name,
int *type_flags_ptr)
{
if (strcmp (name, "short") == 0)
{
*type_flags_ptr = TYPE_FLAG_ADDRESS_CLASS_1;
return 1;
}
else
return 0;
}
The qualifier `@short' is used in GDB's type expressions to indicate
the presence of one of these "short" pointers. For example if the
debug information indicates that `short_ptr_var' is one of these short
pointers, GDB might show the following behavior:
(gdb) ptype short_ptr_var
type = int * @short
File: gdbint.info, Node: Register Representation, Next: Frame Interpretation, Prev: Address Classes, Up: Target Architecture Definition
11.6 Register Representation
============================
* Menu:
* Raw and Cooked Registers::
* Register Architecture Functions & Variables::
* Register Information Functions::
* Register and Memory Data::
* Register Caching::
File: gdbint.info, Node: Raw and Cooked Registers, Next: Register Architecture Functions & Variables, Up: Register Representation
11.6.1 Raw and Cooked Registers
-------------------------------
GDB considers registers to be a set with members numbered linearly from
0 upwards. The first part of that set corresponds to real physical
registers, the second part to any "pseudo-registers". Pseudo-registers
have no independent physical existence, but are useful representations
of information within the architecture. For example the OpenRISC 1000
architecture has up to 32 general purpose registers, which are
typically represented as 32-bit (or 64-bit) integers. However the GPRs
are also used as operands to the floating point operations, and it
could be convenient to define a set of pseudo-registers, to show the
GPRs represented as floating point values.
For any architecture, the implementer will decide on a mapping from
hardware to GDB register numbers. The registers corresponding to real
hardware are referred to as "raw" registers, the remaining registers are
"pseudo-registers". The total register set (raw and pseudo) is called
the "cooked" register set.
File: gdbint.info, Node: Register Architecture Functions & Variables, Next: Register Information Functions, Prev: Raw and Cooked Registers, Up: Register Representation
11.6.2 Functions and Variables Specifying the Register Architecture
-------------------------------------------------------------------
These `struct gdbarch' functions and variables specify the number and
type of registers in the architecture.
-- Architecture Function: CORE_ADDR read_pc (struct regcache *REGCACHE)
-- Architecture Function: void write_pc (struct regcache *REGCACHE,
CORE_ADDR VAL)
Read or write the program counter. The default value of both
functions is `NULL' (no function available). If the program
counter is just an ordinary register, it can be specified in
`struct gdbarch' instead (see `pc_regnum' below) and it will be
read or written using the standard routines to access registers.
This function need only be specified if the program counter is not
an ordinary register.
Any register information can be obtained using the supplied
register cache, REGCACHE. *Note Register Caching: Register
Caching.
-- Architecture Function: void pseudo_register_read (struct gdbarch
*GDBARCH, struct regcache *REGCACHE, int REGNUM, const
gdb_byte *BUF)
-- Architecture Function: void pseudo_register_write (struct gdbarch
*GDBARCH, struct regcache *REGCACHE, int REGNUM, const
gdb_byte *BUF)
These functions should be defined if there are any
pseudo-registers. The default value is `NULL'. REGNUM is the
number of the register to read or write (which will be a "cooked"
register number) and BUF is the buffer where the value read will be
placed, or from which the value to be written will be taken. The
value in the buffer may be converted to or from a signed or
unsigned integral value using one of the utility functions (*note
Using Different Register and Memory Data Representations: Register
and Memory Data.).
The access should be for the specified architecture, GDBARCH. Any
register information can be obtained using the supplied register
cache, REGCACHE. *Note Register Caching: Register Caching.
-- Architecture Variable: int sp_regnum
This specifies the register holding the stack pointer, which may
be a raw or pseudo-register. It defaults to -1 (not defined), but
it is an error for it not to be defined.
The value of the stack pointer register can be accessed withing
GDB as the variable `$sp'.
-- Architecture Variable: int pc_regnum
This specifies the register holding the program counter, which may
be a raw or pseudo-register. It defaults to -1 (not defined). If
`pc_regnum' is not defined, then the functions `read_pc' and
`write_pc' (see above) must be defined.
The value of the program counter (whether defined as a register, or
through `read_pc' and `write_pc') can be accessed withing GDB as
the variable `$pc'.
-- Architecture Variable: int ps_regnum
This specifies the register holding the processor status (often
called the status register), which may be a raw or
pseudo-register. It defaults to -1 (not defined).
If defined, the value of this register can be accessed withing GDB
as the variable `$ps'.
-- Architecture Variable: int fp0_regnum
This specifies the first floating point register. It defaults to
0. `fp0_regnum' is not needed unless the target offers support
for floating point.
File: gdbint.info, Node: Register Information Functions, Next: Register and Memory Data, Prev: Register Architecture Functions & Variables, Up: Register Representation
11.6.3 Functions Giving Register Information
--------------------------------------------
These functions return information about registers.
-- Architecture Function: const char * register_name (struct gdbarch
*GDBARCH, int REGNUM)
This function should convert a register number (raw or pseudo) to a
register name (as a C `const char *'). This is used both to
determine the name of a register for output and to work out the
meaning of any register names used as input. The function may
also return `NULL', to indicate that REGNUM is not a valid
register.
For example with the OpenRISC 1000, GDB registers 0-31 are the
General Purpose Registers, register 32 is the program counter and
register 33 is the supervision register (i.e. the processor status
register), which map to the strings `"gpr00"' through `"gpr31"',
`"pc"' and `"sr"' respectively. This means that the GDB command
`print $gpr5' should print the value of the OR1K general purpose
register 5(1).
The default value for this function is `NULL', meaning undefined.
It should always be defined.
The access should be for the specified architecture, GDBARCH.
-- Architecture Function: struct type * register_type (struct gdbarch
*GDBARCH, int REGNUM)
Given a register number, this function identifies the type of data
it may be holding, specified as a `struct type'. GDB allows
creation of arbitrary types, but a number of built in types are
provided (`builtin_type_void', `builtin_type_int32' etc), together
with functions to derive types from these.
Typically the program counter will have a type of "pointer to
function" (it points to code), the frame pointer and stack pointer
will have types of "pointer to void" (they point to data on the
stack) and all other integer registers will have a type of 32-bit
integer or 64-bit integer.
This information guides the formatting when displaying register
information. The default value is `NULL' meaning no information is
available to guide formatting when displaying registers.
-- Architecture Function: void print_registers_info (struct gdbarch
*GDBARCH, struct ui_file *FILE, struct frame_info *FRAME, int
REGNUM, int ALL)
Define this function to print out one or all of the registers for
the GDB `info registers' command. The default value is the
function `default_print_registers_info', which uses the register
type information (see `register_type' above) to determine how each
register should be printed. Define a custom version of this
function for fuller control over how the registers are displayed.
The access should be for the specified architecture, GDBARCH, with
output to the the file specified by the User Interface Independent
Output file handle, FILE (*note UI-Independent Output--the
`ui_out' Functions: UI-Independent Output.).
The registers should show their values in the frame specified by
FRAME. If REGNUM is -1 and ALL is zero, then all the
"significant" registers should be shown (the implementer should
decide which registers are "significant"). Otherwise only the
value of the register specified by REGNUM should be output. If
REGNUM is -1 and ALL is non-zero (true), then the value of all
registers should be shown.
By default `default_print_registers_info' prints one register per
line, and if ALL is zero omits floating-point registers.
-- Architecture Function: void print_float_info (struct gdbarch
*GDBARCH, struct ui_file *FILE, struct frame_info *FRAME,
const char *ARGS)
Define this function to provide output about the floating point
unit and registers for the GDB `info float' command respectively.
The default value is `NULL' (not defined), meaning no information
will be provided.
The GDBARCH and FILE and FRAME arguments have the same meaning as
in the `print_registers_info' function above. The string ARGS
contains any supplementary arguments to the `info float' command.
Define this function if the target supports floating point
operations.
-- Architecture Function: void print_vector_info (struct gdbarch
*GDBARCH, struct ui_file *FILE, struct frame_info *FRAME,
const char *ARGS)
Define this function to provide output about the vector unit and
registers for the GDB `info vector' command respectively. The
default value is `NULL' (not defined), meaning no information will
be provided.
The GDBARCH, FILE and FRAME arguments have the same meaning as in
the `print_registers_info' function above. The string ARGS
contains any supplementary arguments to the `info vector' command.
Define this function if the target supports vector operations.
-- Architecture Function: int register_reggroup_p (struct gdbarch
*GDBARCH, int REGNUM, struct reggroup *GROUP)
GDB groups registers into different categories (general, vector,
floating point etc). This function, given a register, REGNUM, and
group, GROUP, returns 1 (true) if the register is in the group and
0 (false) otherwise.
The information should be for the specified architecture, GDBARCH
The default value is the function `default_register_reggroup_p'
which will do a reasonable job based on the type of the register
(see the function `register_type' above), with groups for general
purpose registers, floating point registers, vector registers and
raw (i.e not pseudo) registers.
---------- Footnotes ----------
(1) Historically, GDB always had a concept of a frame pointer
register, which could be accessed via the GDB variable, `$fp'. That
concept is now deprecated, recognizing that not all architectures have
a frame pointer. However if an architecture does have a frame pointer
register, and defines a register or pseudo-register with the name
`"fp"', then that register will be used as the value of the `$fp'
variable.
File: gdbint.info, Node: Register and Memory Data, Next: Register Caching, Prev: Register Information Functions, Up: Register Representation
11.6.4 Using Different Register and Memory Data Representations
---------------------------------------------------------------
Some architectures have different representations of data objects,
depending whether the object is held in a register or memory. For
example:
* The Alpha architecture can represent 32 bit integer values in
floating-point registers.
* The x86 architecture supports 80-bit floating-point registers. The
`long double' data type occupies 96 bits in memory but only 80
bits when stored in a register.
In general, the register representation of a data type is determined
by the architecture, or GDB's interface to the architecture, while the
memory representation is determined by the Application Binary Interface.
For almost all data types on almost all architectures, the two
representations are identical, and no special handling is needed.
However, they do occasionally differ. An architecture may define the
following `struct gdbarch' functions to request conversions between the
register and memory representations of a data type:
-- Architecture Function: int gdbarch_convert_register_p (struct
gdbarch *GDBARCH, int REG)
Return non-zero (true) if the representation of a data value
stored in this register may be different to the representation of
that same data value when stored in memory. The default value is
`NULL' (undefined).
If this function is defined and returns non-zero, the `struct
gdbarch' functions `gdbarch_register_to_value' and
`gdbarch_value_to_register' (see below) should be used to perform
any necessary conversion.
If defined, this function should return zero for the register's
native type, when no conversion is necessary.
-- Architecture Function: void gdbarch_register_to_value (struct
gdbarch *GDBARCH, int REG, struct type *TYPE, char *FROM,
char *TO)
Convert the value of register number REG to a data object of type
TYPE. The buffer at FROM holds the register's value in raw
format; the converted value should be placed in the buffer at TO.
_Note:_ `gdbarch_register_to_value' and
`gdbarch_value_to_register' take their REG and TYPE arguments
in different orders.
`gdbarch_register_to_value' should only be used with registers for
which the `gdbarch_convert_register_p' function returns a non-zero
value.
-- Architecture Function: void gdbarch_value_to_register (struct
gdbarch *GDBARCH, struct type *TYPE, int REG, char *FROM,
char *TO)
Convert a data value of type TYPE to register number REG' raw
format.
_Note:_ `gdbarch_register_to_value' and
`gdbarch_value_to_register' take their REG and TYPE arguments
in different orders.
`gdbarch_value_to_register' should only be used with registers for
which the `gdbarch_convert_register_p' function returns a non-zero
value.
File: gdbint.info, Node: Register Caching, Prev: Register and Memory Data, Up: Register Representation
11.6.5 Register Caching
-----------------------
Caching of registers is used, so that the target does not need to be
accessed and reanalyzed multiple times for each register in
circumstances where the register value cannot have changed.
GDB provides `struct regcache', associated with a particular `struct
gdbarch' to hold the cached values of the raw registers. A set of
functions is provided to access both the raw registers (with `raw' in
their name) and the full set of cooked registers (with `cooked' in
their name). Functions are provided to ensure the register cache is
kept synchronized with the values of the actual registers in the target.
Accessing registers through the `struct regcache' routines will
ensure that the appropriate `struct gdbarch' functions are called when
necessary to access the underlying target architecture. In general
users should use the "cooked" functions, since these will map to the
"raw" functions automatically as appropriate.
The two key functions are `regcache_cooked_read' and
`regcache_cooked_write' which read or write a register from or to a
byte buffer (type `gdb_byte *'). For convenience the wrapper functions
`regcache_cooked_read_signed', `regcache_cooked_read_unsigned',
`regcache_cooked_write_signed' and `regcache_cooked_write_unsigned' are
provided, which read or write the value using the buffer and convert to
or from an integral value as appropriate.
File: gdbint.info, Node: Frame Interpretation, Next: Inferior Call Setup, Prev: Register Representation, Up: Target Architecture Definition
11.7 Frame Interpretation
=========================
* Menu:
* All About Stack Frames::
* Frame Handling Terminology::
* Prologue Caches::
* Functions and Variable to Analyze Frames::
* Functions to Access Frame Data::
* Analyzing Stacks---Frame Sniffers::
File: gdbint.info, Node: All About Stack Frames, Next: Frame Handling Terminology, Up: Frame Interpretation
11.7.1 All About Stack Frames
-----------------------------
GDB needs to understand the stack on which local (automatic) variables
are stored. The area of the stack containing all the local variables
for a function invocation is known as the "stack frame" for that
function (or colloquially just as the "frame"). In turn the function
that called the function will have its stack frame, and so on back
through the chain of functions that have been called.
Almost all architectures have one register dedicated to point to the
end of the stack (the "stack pointer"). Many have a second register
which points to the start of the currently active stack frame (the
"frame pointer"). The specific arrangements for an architecture are a
key part of the ABI.
A diagram helps to explain this. Here is a simple program to compute
factorials:
#include <stdio.h>
int fact (int n)
{
if (0 == n)
{
return 1;
}
else
{
return n * fact (n - 1);
}
}
main ()
{
int i;
for (i = 0; i < 10; i++)
{
int f = fact (i);
printf ("%d! = %d\n", i, f);
}
}
Consider the state of the stack when the code reaches line 6 after
the main program has called `fact (3)'. The chain of function calls
will be `main ()', `fact (3)', `fact (2)', `fact (1)' and `fact (0)'.
In this illustration the stack is falling (as used for example by the
OpenRISC 1000 ABI). The stack pointer (SP) is at the end of the stack
(lowest address) and the frame pointer (FP) is at the highest address
in the current stack frame. The following diagram shows how the stack
looks.