dmalloc.texi

\input texinfo @c -*-texinfo-*-
@c %**start of header
@setfilename dmalloc.info
@settitle Dmalloc Tutorial
@setchapternewpage odd
@headings double
@c %**end of header
  
@dircategory Libraries
@direntry
* Dmalloc: (dmalloc).			Malloc debug library.
@end direntry

@c ----------------------------------------------------------------
@set dmalloc_version Version 5.5.2
@set dmalloc_date May 2007
@set dmalloc_copyright Copyright 1992 to 2007 by Gray Watson.
@c ----------------------------------------------------------------

@ifinfo
This file is an introduction to the Dmalloc library which handles general
memory heap management.

@value{dmalloc_copyright}

Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

@ignore
Permission is granted to process this file through TeX and print the
results, provided the printed document carries a copying permission
notice identical to this one except for the removal of this paragraph
(this paragraph not being relevant to the printed manual).

@end ignore
Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided also that the
chapter entitled ``Copying'' are included exactly as in the original,
and provided that the entire resulting derived work is distributed under
the terms of a permission notice identical to this one.

Permission is granted to copy and distribute translations of this manual
into another language, under the above conditions for modified versions,
except that the chapter entitled ``Copying'' may be included in a
translation approved by the author instead of in the original English.
@end ifinfo

@titlepage
@title Debug Malloc Library
@subtitle @value{dmalloc_version}
@subtitle @value{dmalloc_date}
@author Gray Watson

@page
@value{dmalloc_copyright}

Published by Gray Watson

Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided also that the
chapter entitled ``Copying'' are included exactly as in the original,
and provided that the entire resulting derived work is distributed under
the terms of a permission notice identical to this one.

Permission is granted to copy and distribute translations of this manual
into another language, under the above conditions for modified versions,
except that the chapter entitled ``Copying'' may be included in a
translation approved by the author instead of in the original English.
@end titlepage

@c --------------------------------

@node Top, Copying, (dir), (dir)
@top Debug Malloc Library

@ifinfo
@value{dmalloc_version} -- @value{dmalloc_date}
@end ifinfo

@cindex introduction
@cindex author

The debug memory allocation or @dfn{dmalloc} library has been designed
as a drop in replacement for the system's @code{malloc}, @code{realloc},
@code{calloc}, @code{free} and other memory management routines while
providing powerful debugging facilities configurable at runtime.  These
facilities include such things as memory-leak tracking, fence-post write
detection, file/line number reporting, and general logging of
statistics.

The library is reasonably portable having been run successfully on at
least the following operating systems: AIX, DGUX, Free/Net/OpenBSD,
GNU/Hurd, HPUX, Irix, Linux, Mac OSX, NeXT, OSF/DUX, SCO, Solaris,
Ultrix, Unixware, MS Windows, and Unicos on a Cray T3E.  It also
provides support for the debugging of threaded programs.  @xref{Using
With Threads}.

The package includes the library, configuration scripts, debug utility
application, test program, and documentation.  Online documentation as
well as the full source is available at URL @uref{http://dmalloc.com/}.
Details on the library's mailing list are available there as well.

Please use the forums at URL @uref{http://dmalloc.com/} to discuss any
problems or to request features.  If you are still having problems,
the author can be reached via his home page at URL
@uref{http://256.com/gray/} with questions or feedback.  Please
include the version number of the library that you are using, your
machine and operating system types, and the value of the
DMALLOC_OPTIONS environment variable.

Gray Watson.

@menu
* Copying::                     Library copying and licensing conditions.
* Overview::                    Description of features and how to get started.
* Programming::                 How to program with the library.
* Dmalloc Program::             How to use the library's utility.
* Source Code::                 Information on the source code.
* Troubleshooting::             Some solutions to common problems.
* Index of Concepts::           Index of concepts in the manual.
@end menu

@c ----------------------------------------------------------------------------

@node Copying, Overview, Top, Top
@chapter Library Copying and Licensing Conditions

@cindex copying
@cindex license
@cindex library permissions
@cindex permissions of the library

@value{dmalloc_copyright}

Permission to use, copy, modify, and distribute this software for any
purpose and without fee is hereby granted, provided that the above
copyright notice and this permission notice appear in all copies, and
that the name of Gray Watson not be used in advertising or publicity
pertaining to distribution of the document or software without specific,
written prior permission.

Gray Watson makes no representations about the suitability of the
software described herein for any purpose.  It is provided ``as is''
without express or implied warranty.

@c ----------------------------------------------------------------------------

@node Overview, Programming, Copying, Top
@chapter Description of Features and How to Get Started

@cindex overview

@menu
* Installation::                How to install the library.
* Getting Started::             Getting started with the library.
* Allocation Basics::           Basic description of terms and functions.
* Features::                    General features of the library.
* How It Works::                How the library checks your program.
@end menu

@c --------------------------------

@node Installation, Getting Started, Overview, Overview
@section How to Install the Library

@cindex installing the library
@cindex compiling the library
@cindex building the library
@cindex configuring the library
@cindex making the library

To configure, compile, and install the library, follow these steps
carefully.

@enumerate

@item Make sure you have downloaded the latest version of the library
available from the home page at URL @uref{http://dmalloc.com/}.

@item The release files have a @file{.tgz} file extension which means
that they are a tar'd gzip'd directory of files.  You will need to
ungzip and then untar the release file into your source work
directory.  You may have to rename the file to @file{.tar.gz} to get
some old zip programs to handle the file correctly.

@cindex settings.dist file

@item You may want to edit or at least review the settings in
@file{settings.dist} to tune specific features of the library.  The
@file{configure} script will copy this file to @file{settings.h} which
is where you should be adding per-architecture settings.

@cindex configure script
@cindex --disable-cxx
@cindex --enable-threads
@cindex --enable-shlib
@cindex conf.h file

@item Type @kbd{sh ./configure} to configure the library.  You may want
to first examine the @file{config.help} file for some information about
configure.  You may want to use the @kbd{--disable-cxx} option if you do
not want the Makefile to build the C++ version of dmalloc.  You may want
to use the @kbd{--enable-threads} option to build the threaded version
of dmalloc.  You may want to use the @kbd{--enable-shlib} option to
build the shared versions of the dmalloc libraries.  @kbd{sh ./configure
--help} lists the available options to configure.  Configure should
generate the @file{Makefile} and configuration files automatically.

@item You may want to examine the @file{Makefile} and @file{conf.h} files
created by configure to make sure it did its job correctly.

@cindex settings.h file

@item You might want to tune the settings in @file{settings.h} file to
tune the library to the local architecture.  This file contains relevant
settings if you are using pthreads or another thread library.
@xref{Using With Threads}.  The @file{configure} script created this
file from the @file{settings.dist} file.  Any permanent changes to these
settings should made to the @file{settings.dist} file.  You then can run
@file{config.status} to re-create the @file{settings.h} file.

@cindex DMALLOC_SIZE option

@item The @code{DMALLOC_SIZE} variable gets auto-configured in
@file{dmalloc.h.2} but it may not generate correct settings for all
systems.  You may have to alter the definitions in this file to get
things to stop complaining when you go to compile about the size
arguments to malloc routines.  Comments on this please.

@item Typing @kbd{make} should be enough to build @file{libdmalloc.a},
and @file{dmalloc} program.  If it does not work, please see if there
are any notes in the contrib directory about your system-type.  If not
and you figure your problem out, please send me some notes so future
users can profit from your experiences.

@cindex USE_RETURN_MACROS conf.h option

@emph{NOTE}: You may experience some errors compiling some of the
@file{return.h} assembly macros which attempt to determine the callers
address for logging purposes.  @xref{Portability}.  You may want to
first try disabling any compiler optimization flags.  If this doesn't
work then you may need to disable the @samp{USE_RETURN_MACROS}
variable in the @file{settings.h} file.

@cindex ANSI-C compiler

@emph{NOTE}: The code is dependent on an ANSI-C compiler.  If the
configure script gives the @samp{WARNING} that you do not have an ANSI-C
compiler, you may still be able to add some sort of option to your
compiler to make it ANSI.  If there such is an option, please send it to
the author so it can be added to the configure script.

@item If you use threads and did not add the @kbd{--enable-threads}
argument to configure, typing @kbd{make threads} should be enough to
build @file{libdmallocth.a} which is the threaded version of the
library.  This may or may not work depending on the configuration
scripts ability to detect your local thread functionality.  Feel free to
send me mail with improvements.

See the section of the manual on threads for more information about
the operation of the library with your threaded program.  @xref{Using
With Threads}.

@item If you have a C++ compiler installed, the library should have
automatically built @file{libdmallocxx.a} which is the C++ version of
the library.  If it was not done automatically, you can build it by
typing @kbd{make cxx}.  You should link this library into your C++
programs instead of @file{libdmalloc.a}.  See the @file{dmallocc.cc} C++
file which contains basic code to overload the @code{new}, @code{new[]},
@code{delete}, and @code{delete[]} C++ operators.  My apologies on the
minimal C++ support.  I am still living in a mostly C world.  Any help
improving this interface without sacrificing portability would be
appreciated.

@cindex testing the library
@cindex dmalloc_t test program

@item Typing @kbd{make light} should build and run the @file{dmalloc_t} test
program through a set of light trials.  By default this will execute
@file{dmalloc_t} 5 times -- each time will execute 10,000 malloc
operations in a very random manner.  Anal folks can type @kbd{make
heavy} to up the ante.  Use @kbd{dmalloc_t --usage} for the list of all
@file{dmalloc_t} options.

@item Typing @kbd{make install} should install the @file{libdmalloc.a}
library in @file{/usr/local/lib}, the @file{dmalloc.h} include file in
@file{/usr/local/include}, and the @file{dmalloc} utility in
@file{/usr/local/bin}.  You may also want to type @kbd{make installth}
to install the thread library into place and/or @kbd{make installcc} to
install the C++ library into place.

You may have specified a @samp{--prefix=PATH} option to configure in
which case @samp{/usr/local} will have been replaced with @samp{PATH}.

@end enumerate

See the ``Getting Started'' section to get up and running with the
library.  @xref{Getting Started}.

@c --------------------------------

@node Getting Started, Allocation Basics, Installation, Overview
@section Getting Started with the Library

@cindex quick start
@cindex getting started
@cindex jump start
@cindex how to begin
@cindex where to begin
@cindex beginning

This section should give you a quick idea on how to get going.
Basically, you need to do the following things to make use of the
library:

@enumerate

@item Make sure you have downloaded the latest version of the library
available from the home page at URL @uref{http://dmalloc.com/}.

@item Follow the installation instructions on how to configure,
make, and install the library (i.e. type: @kbd{make install}).
@xref{Installation}.

@cindex automatic shutdown
@cindex shutdown, automatic
@cindex on_exit
@cindex atexit
@cindex destructor

@item You need to make sure that the library configuration and
build process above was able to locate one of the @code{on_exit}
function, @code{atexit} function, or had compiler destructor support.
If one of these functions or support is available then the dmalloc
library should be able to automatically shut itself down when the
program exits.  This causes the memory statistics and unfreed
information to be dumped to the log file.  However, if your system has
none of the above, then you will need to call @code{dmalloc_shutdown}
yourself before your program exits.

@cindex alias, shell
@cindex bash shell
@cindex ksh shell
@cindex zsh shell

@item To get the dmalloc utility to work you need to add an alias for
dmalloc to your shell's runtime configuration file if supported.  The
idea is to have the shell capture the dmalloc program's output and
adjust the environment.  

After you add the alias to the shell config file you need to log out
and log back in to have it take effect, or you can execute the
appropriate command below on the command line directly.  After you
setup the alias, if you enter @kbd{dmalloc runtime} and see any output
with DMALLOC_OPTIONS in it then the alias did not take effect.

Bash, ksh, and zsh (@uref{http://www.zsh.org/}) users should add the
following to their @file{.bashrc}, @file{.profile}, or @file{.zshrc}
file respectively (notice the @kbd{-b} option for bourne shell
output):

@example
function dmalloc @{ eval `command dmalloc -b $*`; @}
@end example

If your shell does not support the @code{command} function then try:

@example
function dmalloc @{ eval `\dmalloc -b $*`; @}
@end example
or
@example
function dmalloc @{ eval `/usr/local/bin/dmalloc -b $*`; @}
@end example

@cindex csh shell
@cindex tcsh shell

If you are still using csh or tcsh, you should add the following to
your @file{.cshrc} file (notice the @kbd{-C} option for c-shell
output):

@example
alias dmalloc 'eval `\dmalloc -C \!*`'
@end example

@cindex rc shell

If you are using rc shell, you should add the following to your
@file{.rcrc} file (notice the @kbd{-R} option for rc-shell output):

@example
fn dmalloc @{eval `@{/usr/local/bin/dmalloc $*@}@}
@end example

@item Although not necessary, you may want to include @file{dmalloc.h}
in your C files and recompile.  This will allow the library to report
the file/line numbers of calls that generate problems.  @xref{Allocation
Macros}.  It should be inserted at the @emph{bottom} of your include
files as to not conflict with wother includes.  You may want to ifdef it
as well and compile with @kbd{cc -DDMALLOC @dots{}}:

@example
/* other includes above ^^^ */

#ifdef DMALLOC
#include "dmalloc.h"
#endif
@end example

@cindex DMALLOC_FUNC_CHECK
@item Another optional task is to compile all of your source with the
@file{dmalloc.h} with the @code{DMALLOC_FUNC_CHECK} compilation flag.
This willallow the library to check all of the arguments of a number
of common string and utility routines.  @xref{Argument Checking}.

@example
cc -DDMALLOC -DDMALLOC_FUNC_CHECK file.c
@end example

@item Link the dmalloc library into your program.  The dmalloc library
should probably be placed at or near the end of the library list.

@item Enable the debugging features by typing @kbd{dmalloc -l logfile
-i 100 low} (for example).  You should not see any messages printed by
the dmalloc utility (see NOTE below).  This will:

@itemize @bullet

@item Set the malloc logfile name to @file{logfile} (@kbd{-l logfile}).
For programs which change directories, you may want to specify the
full path to your logfile.

@item Have the library check itself every 100 iterations (@kbd{-i 100}).
This controls how fast your program will run.  Larger numbers check
the heap less and so it will run faster.  Lower numbers will be more
likely to catch memory problems.

@item Enable a number of debug features (@kbd{low}).  You can
also try @kbd{runtime} for minimal checking or @kbd{medium} or
@kbd{high} for more extensive heap verification.

@item By default, the low, medium, and high values enable the
@code{error-abort} token which will cause the library to abort and
usually dump core immediately upon seeing an error.  @xref{Dumping
Core}.  You can disable this feature by entering @kbd{dmalloc -m
error-abort} (-m for minus) to remove the @code{error-abort} token and
your program will just log errors and continue.

@end itemize

@kbd{dmalloc --usage} will provide verbose usage info for the dmalloc
program.  @xref{Dmalloc Program}.

You may also want to install the @file{dmallocrc} file in your home
directory as @file{.dmallocrc}.  This allows you to add your own
combination of debug tokens.  @xref{RC File}.

@emph{NOTE}: The output from the dmalloc utility should be captured by
your shell.  If you see a bunch of stuff which includes the string
@code{DMALLOC_OPTIONS} then the alias you should have created above is
not working and he environmental variables are not being set.  Make
sure you've logged out and back in to have the alias take effect.

@item Run your program, examine the logfile that should have been
created by @code{dmalloc_shutdown}, and use its information to help
debug your program.

@end enumerate

@c --------------------------------

@node Allocation Basics, Features, Getting Started, Overview
@section Basic Description of Terms and Functions

@cindex allocation basics
@cindex basic allocation information

@menu
* Basic Definitions::           General memory terms and concepts.
* Malloc Functions::            Functionality supported by all malloc libs.
@end menu

@c --------------------------------

@node Basic Definitions, Malloc Functions, Allocation Basics, Allocation Basics
@subsection General Memory Terms and Concepts

@cindex basic definitions
@cindex memory definitions

Any program can be divided into 2 logical parts: text and data.  Text is
the actual program code in machine-readable format and data is the
information that the text operates on when it is executing.  The data,
in turn, can be divided into 3 logical parts according to where it is
stored: @dfn{static}, @dfn{stack}, and @dfn{heap}.

@cindex static memory

Static data is the information whose storage space is compiled into the
program.

@example
/* global variables are allocated as static data */
int numbers[10];

main()
@{
        @dots{}
@}
@end example

@cindex stack memory

Stack data is data allocated at runtime to hold information used inside
of functions.  This data is managed by the system in the space called
stack space.

@example
void foo()
@{
        /* this local variable is stored on the stack */
        float total;
        @dots{}
@}

main()
@{
        foo();
@}
@end example

@cindex heap memory

Heap data is also allocated at runtime and provides a programmer with
dynamic memory capabilities.

@example
main()
@{
        /* the address is stored on the stack */
        char * string;
        @dots{}

        /*
         * Allocate a string of 10 bytes on the heap.  Store the
         * address in string which is on the stack.
         */
        string = (char *)malloc(10);
        @dots{}

        /* de-allocate the heap memory now that we're done with it */
        (void)free(string);
        @dots{}
@}
@end example

It is the heap data that is managed by this library.

Although the above is an example of how to use the malloc and free
commands, it is not a good example of why using the heap for runtime
storage is useful.

Consider this: You write a program that reads a file into memory,
processes it, and displays results.  You would like to handle files with
arbitrary size (from 10 bytes to 1.2 megabytes and more).  One problem,
however, is that the entire file must be in memory at one time to do the
calculations.  You don't want to have to allocate 1.2 megabytes when you
might only be reading in a 10 byte file because it is wasteful of system
resources.  Also, you are worried that your program might have to handle
files of more than 1.2 megabytes.

A solution: first check out the file's size and then, using the
heap-allocation routines, get enough storage to read the entire file
into memory.  The program will only be using the system resources
necessary for the job and you will be guaranteed that your program can
handle any sized file.

@c --------------------------------

@node Malloc Functions,, Basic Definitions, Allocation Basics
@subsection Functionality Supported by All Malloc Libraries

@cindex malloc functions

All malloc libraries support 4 basic memory allocation commands.  These
include @dfn{malloc}, @dfn{calloc}, @dfn{realloc}, and @dfn{free}.  For
more information about their capabilities, check your system's manual
pages -- in unix, do a @code{man 3 malloc}.

@cindex malloc
@deftypefun void *malloc ( unsigned int @var{size} )

Usage: @code{pnt = (type *)malloc(size)}

The malloc routine is the basic memory allocation routine.  It allocates
an area of @code{size} bytes.  It will return a pointer to the space
requested.

@end deftypefun
@deftypefun void *calloc ( unsigned int @var{number}, unsigned int@var{size} )
@cindex calloc
@cindex Allocation of zeros
@cindex zeros, allocation of

Usage: @code{pnt = (type *)calloc(number, size)}

The calloc routine allocates a certain @code{number} of items, each of
@code{size} bytes, and returns a pointer to the space.  It is
appropriate to pass in a @code{sizeof(type)} value as the size argument.

Also, calloc nulls the space that it returns, assuring that the memory
is all zeros.

@end deftypefun
@deftypefun void *realloc ( void *@var{old_pnt}, unsigned int @var{new_size} )
@cindex realloc

Usage: @code{new_pnt = (type *)realloc(old_pnt, new_size)}

The realloc function expands or shrinks the memory allocation in
@code{old_pnt} to @code{new_size} number of bytes.  Realloc copies as
much of the information from @code{old_pnt} as it can into the
@code{new_pnt} space it returns, up to @code{new_size} bytes.  If there
is a problem allocating this memory, 0L will be returned.

If the @code{old_pnt} is 0L then realloc will do the equivalent of a
@code{malloc(new_size)}.  If @code{new_size} is 0 and @code{old_pnt} is
not 0L, then it will do the equivalent of @code{free(old_pnt)} and will
return 0L.

@end deftypefun
@deftypefun void free ( void *@var{pnt} )
@cindex free

Usage: @code{free(pnt)}

The free routine releases allocation in @code{pnt} which was returned by
malloc, calloc, or realloc back to the heap.  This allows other parts of
the program to re-use memory that is not needed anymore.  It guarantees
that the process does not grow too big and swallow a large portion of
the system resources.

@end deftypefun

@emph{WARNING}: there is a quite common myth that all of the space that
is returned by malloc libraries has already been cleared.  @emph{Only}
the @code{calloc} routine will zero the memory space it returns.

@c --------------------------------

@node Features, How It Works, Allocation Basics, Overview
@section General Features of the Library

@cindex features

The debugging features that are available in this debug malloc library
can be divided into a couple basic classifications:

@table @asis
@item file and line number information
@cindex file/line numbers
@cindex cpp
One of the nice things about a good debugger is its ability to provide
the file and line number of an offending piece of code.  This library
attempts to give this functionality with the help of @dfn{cpp}, the C
preprocessor.  @xref{Allocation Macros}.

@item return-address information
@cindex return-address
To debug calls to the library from external sources (i.e. those files
that could not use the allocation macros), some facilities have been
provided to supply the caller's address.  This address, with the help of
a debugger, can help you locate the source of a problem.  @xref{Return
Address}.

@item fence-post (i.e. bounds) checking
@cindex fence-post checking
@cindex bounds checking
@cindex checking bounds
@dfn{Fence-post} memory is the area immediately above or below memory
allocations.  It is all too easy to write code that accesses above or
below an allocation -- especially when dealing with arrays or strings.
The library can write special values in the areas around every
allocation so it will notice when these areas have been overwritten.
@xref{Fence-Post Overruns}.

@emph{NOTE}: The library cannot notice when the program @emph{reads}
from these areas, only when it writes values.  Also, fence-post checking
will increase the amount of memory the program allocates.

@item heap-constancy verification
@cindex constancy verification
The administration of the library is reasonably complex.  If any of the
heap-maintenance information is corrupted, the program will either crash
or give unpredictable results.

By enabling heap-consistency checking, the library will run through its
administrative structures to make sure all is in order.  This will mean
that problems will be caught faster and diagnosed better.

The drawback of this is, of course, that the library often takes quite a
long time to do this.  It is suitable to enable this only during
development and debugging sessions.

@emph{NOTE}: the heap checking routines cannot guarantee that the tests
will not cause a segmentation-fault if the heap administration
structures are properly (or improperly if you will) overwritten.  In
other words, the tests will verify that everything is okay but may not
inform the user of problems in a graceful manner.

@item logging statistics
@cindex logging statistics
@cindex statistics
@cindex memory leaks
@cindex leaking memory
One of the reasons why the debug malloc library was initially developed
was to track programs' memory usage -- specifically to locate
@dfn{memory leaks} which are places where allocated memory is never
getting freed.  @xref{Memory Leaks}.

The library has a number of logging capabilities that can track un-freed
memory pointers as well as runtime memory usage, memory transactions,
administrative actions, and final statistics.

@item examining freed memory
@cindex freed memory
@cindex freed memory
Another common problem happens when a program frees a memory pointer but
goes on to use it again by mistake.  This can lead to mysterious crashes
and unexplained problems.

To combat this, the library can write special values into a block of
memory after it has been freed.  This serves two purposes: it will make
sure that the program will get garbage data if it trying to access the
area again, and it will allow the library to verify the area later for
signs of overwriting.
@end table

If any of the above debugging features detect an error, the library will
try to recover.  If logging is enabled then an error will be logged with
as much information as possible.

The error messages that the library displays are designed to give the
most information for developers.  If the error message is not
understood, then it is most likely just trying to indicate that a part
of the heap has been corrupted.

@cindex dump core
@cindex core dump
The library can be configured to quit immediately when an error is
detected and to dump a core file or memory-image.  This can be
examined with a debugger to determine the source of the problem.  The
library can either stop after dumping core or continue running.
@xref{Dumping Core}.

@cindex system memory problems
@cindex memory problems in system functions
@emph{NOTE}: do not be surprised if the library catches problems with
your system's routines.  It took me hours to finally come to the
conclusion that the localtime call, included in SunOS release 4.1,
overwrites one of its fence-post markers.

@c --------------------------------

@node How It Works,, Features, Overview
@section How the Library Checks Your Program

This is one of the newer sections of the library implying that it is
incomplete.  If you have any questions or issues that you'd like to see
handled here, please let me know.

The dmalloc library replaces the heap library calls normally found in
your system libraries with its own versions.  When you make a call to
malloc (for example), you are calling dmalloc's version of the memory
allocation function.  When you allocate memory with these functions, the
dmalloc library keeps track of a number of pieces of debugging
information about your pointer including: where it was allocated,
exactly how much memory was requested, when the call was made, etc..
This information can then be verified when the pointer is freed or
reallocated and the details can be logged on any errors.

Whenever you reallocate or free a memory address, the dmalloc library
always performs a number of checks on the pointer to make sure that it
is valid and has not been corrupted.  You can configure the library to
perform additional checks such as detected fence-post writing.  The
library can also be configured to overwrite memory with non-zeros (only
if calloc is not called) when it is allocated and erase the memory when
the pointers are freed.

@cindex check-heap
In addition to per-pointer checks, you can configure the library to
perform complete heap checks.  These complete checks verify all
internal heap structures and include walking all of the known
allocated pointers to verify each one in turn.  You need this level of
checking to find random pointers in your program which got corrupted
but that won't be freed for a while.  To turn on these checks, you
will need to enable the @code{check-heap} debug token.  @xref{Debug
Tokens}.  By default this will cause the heap to be fully checked each
and every time dmalloc is called whether it is a malloc, free,
realloc, or another dmalloc overloaded function.

Performing a full heap check can take a good bit of CPU and it may be
that you will want to run it sporadically.  This can be accomplished in
a couple different ways including the '-i' interval argument to the
dmalloc utility.  @xref{Dmalloc Program}.  This will cause the check to
be run every N-th time.  For instance, 'dmalloc -i 3' will cause the
heap to be checked before every 3rd call to a memory function.  Values
of 100 or even 1000 for high memory usage programs are more useful than
smaller ones.

@cindex LOG_ITERATION_COUNT
@cindex iteration count
@cindex memory transaction count
You can also cause the program to start doing detailed heap checking
after a certain point.  For instance, with 'dmalloc -s 1000' option, you
can tell the dmalloc library to enable the heap checks after the 1000th
memory call.  Examine the dmalloc log file produced and use the
iteration count if you have @code{LOG_ITERATION_COUNT} enabled in your
@file{settings.h} file.

The start option can also have the format @samp{file:line}.  For
instance, if it is set to @samp{dmalloc_t.c:126}, dmalloc will start
checking the heap after it sees a dmalloc call from the
@file{dmalloc_t.c} file, line number 126.  If you use
@samp{dmalloc_t.c:0}, with a 0 line number, then dmalloc will start
checking the heap after it sees a call from anywhere in the
@file{dmalloc_t.c} file.

@c ----------------------------------------------------------------------------

@node Programming, Dmalloc Program, Overview, Top
@chapter How to Program with the Library

@cindex programming

@menu
* Allocation Macros::          Macros providing file and line information.
* Return Address::             Getting caller address information.
* Argument Checking::          Checking of function arguments.
* Dumping Core::               Generating a core file on errors for debugging.
* Extensions::                 Additional non-standard routines.
* Error Codes::                Description of the internal error numbers.
* Disabling the Library::      How to disable the library.
* Using With C++::             Using the library with C++.
* Using With a Debugger::      Using a debugger with the library.
* Using With Threads::         Using the library with a thread package.
* Using With Cygwin::          Using the library with Cygwin environment.
* Debugging A Server::         Debugging memory in a server or cgi-bin process.
* Logfile Details::            Explanation of the Logfile Output.
* Other Hints::                Various other hints that may help.
@end menu

@c --------------------------------

@node Allocation Macros, Return Address, Programming, Programming
@section Macros Providing File and Line Information

@cindex allocation macros
@cindex macros, allocation
@cindex dmalloc.h file

By including @file{dmalloc.h} in your C files, your calls to malloc,
calloc, realloc, recalloc, memalign, valloc, strdup, and free are
replaced with calls to _dmalloc_malloc, _dmalloc_realloc, and
_dmalloc_free with various flags.  Additionally the library replaces
calls to xmalloc, xcalloc, xrealloc, xrecalloc, xmemalign, xvalloc,
xstrdup, and xfree with associated calls.

These macros use the c-preprocessor @code{__FILE__} and @code{__LINE__}
macros which get replaced at compilation time with the current file and
line-number of the source code in question.  The routines use this
information to produce verbose reports on memory problems.

@example
not freed: '0x38410' (22 bytes) from 'dmalloc_t.c:92'
@end example

This line from a log file shows that memory was not freed from file
@file{dmalloc_t.c} line 92.  @xref{Memory Leaks}.

@cindex recalloc
@cindex memalign
@cindex valloc
@cindex strdup

You may notice some non standard memory allocation functions in the
above list.  Recalloc is a routine like realloc that reallocates
previously allocated memory to a new size.  If the new memory size is
larger than the old, recalloc initializes the new space to all zeros.
This may or may not be supported natively by your operating system.
Memalign is like malloc but should insure that the returned pointer is
aligned to a certain number of specified bytes.  Currently, the memalign
function is not supported by the library.  It defaults to returning
possibly non-aligned memory for alignment values less than a block-size.
Valloc is like malloc but insures that the returned pointer will be
aligned to a page boundary.  This may or may not be supported natively
by your operating system but is fully supported by the library.  Strdup
is a string duplicating routine which takes in a null terminated string
pointer and returns an allocated copy of the string that will need to be
passed to free later to deallocate.

The X versions of the standard memory functions (xmalloc, xfree, etc.)
will print out an error message to standard error and will stop if the
library is unable to allocate any additional memory.  It is useful to
use these routines instead of checking everywhere in your program for
allocation routines returning NULL pointers.

@emph{WARNING}: If you are including the @file{dmalloc.h} file in your
sources, it is recommended that it be at the end of your include file
list because dmalloc uses macros and may try to change declarations of
the malloc functions if they come after it.

@c --------------------------------

@node Return Address, Argument Checking, Allocation Macros, Programming
@section Getting Caller Address Information

@cindex caller's address
@cindex return-address
@cindex return.h file
@cindex ra

Even though the allocation macros can provide file/line information for
some of your code, there are still modules which either you can't
include @file{dmalloc.h} (such as library routines) or you just don't
want to.  You can still get information about the routines that call
dmalloc function from the return-address information.  To accomplish
this, you must be using this library on one of the supported
architecture/compilers.  @xref{Portability}.

@cindex assembly hacks

The library attempts to use some assembly hacks to get the
return-address or the address of the line that called the dmalloc
function.  If you have unfreed memory that does not have associated file
and line information, you might see the following non-freed memory
messages.

@example
not freed: '0x38410' (22 bytes) from 'ra=0xdd2c'
not freed: '0x38600' (10232 bytes) from 'ra=0x10234d'
not freed: '0x38220' (137 bytes) from 'ra=0x82cc'
@end example

@cindex ra_info.pl
With the help of a debugger, these return-addresses (or ra) can then be
identified.  I've provided a @file{ra_info.pl} perl script in the
@file{contrib/} directory with the dmalloc sources which seems to work
well with gdb.  You can also use manual methods for gdb to find the
return-address location.  @xref{Translate Return Addresses}.

@c --------------------------------

@node Argument Checking, Dumping Core, Return Address, Programming
@section Checking of Function Arguments

@cindex argument checking
@cindex checking arguments
@cindex DMALLOC_FUNC_CHECK flag

One potential problem with the library and its multitude of checks and
diagnoses is that they only get performed when a dmalloc function is
called.  One solution this is to include @file{dmalloc.h} and compile
your source code with the @code{DMALLOC_FUNC_CHECK} flag defined and
enable the @code{check-funcs} token.  @xref{Debug Tokens}.

@example
cc -DDMALLOC -DDMALLOC_FUNC_CHECK file.c
@end example

@emph{NOTE}: Once you have compiled your source with DMALLOC_FUNC_CHECK
enabled, you will have to recompile with it off to disconnect the
library.  @xref{Disabling the Library}.

@emph{WARNING}: You should be sure to have @file{dmalloc.h} included at
the end of your include file list because dmalloc uses macros and may
try to change declarations of the checked functions if they come after
it.

When this is defined dmalloc will override a number of functions and
will insert a routine which knows how to check its own arguments and
then call the real function.  Dmalloc can check such functions as
@code{bcopy}, @code{index}, @code{strcat}, and @code{strcasecmp}.  For
the full list see the end of @file{dmalloc.h}.

When you call @code{strlen}, for instance, dmalloc will make sure the
string argument's fence-post areas have not been overwritten, its file
and line number locations are good, etc.  With @code{bcopy}, dmalloc
will make sure that the destination string has enough space to store the
number of bytes specified.

For all of the arguments checked, if the pointer is not in the heap then
it is ignored since dmalloc does not know anything about it.

@c --------------------------------

@node Dumping Core, Extensions, Argument Checking, Programming
@section Generating a Core File on Errors

@cindex dump core
@cindex core dump
@cindex generating core dump
@cindex error-abort
@cindex error-dump
@cindex debugger

If the @code{error-abort} debug token has been enabled, when the
library detects any problems with the heap memory, it will immediately
attempt to dump a core file.  @xref{Debug Tokens}.  Core files are a
complete copy of the program and it's state and can be used by a
debugger to see specifically what is going on when the error occurred.
@xref{Using With a Debugger}.  By default, the low, medium, and high
arguments to the library utility enable the @code{error-abort} token.
You can disable this feature by entering @kbd{dmalloc -m error-abort}
(-m for minus) to remove the @code{error-abort} token and your program
will just log errors and continue.  You can also use the
@code{error-dump} token which tries to dump core when it sees an error
but still continue running.  @xref{Debug Tokens}.

@cindex core file
@cindex program core
@cindex no core dump
@cindex core dump, none
@cindex missing core dump

When a program dumps core, the system writes the program and all of
its memory to a file on disk usually named @file{core}.  If your
program is called @file{foo} then your system may dump core as
@file{foo.core}.  If you are not getting a @file{core} file, make sure
that your program has not changed to a new directory meaning that it
may have written the core file in a different location.  Also insure
that your program has write privileges over the directory that it is
in otherwise it will not be able to dump a core file.  Core dumps are
often security problems since they contain all program memory so
systems often block their being produced.  You will want to check your
user and system's core dump size ulimit settings.

@cindex abort
@cindex kill process
@cindex KILL_PROCESS

The library by default uses the @code{abort} function to dump core
which may or may not work depending on your operating system.  If the
following program does not dump core then this may be the problem.
See @code{KILL_PROCESS} definition in @file{settings.dist}.

@example
main()
@{
  abort();
@}
@end example

If @code{abort} does work then you may want to try the following
setting in @file{settings.dist}.  This code tries to generate a
segmentation fault by dereferencing a @code{NULL} pointer.

@example
#define KILL_PROCESS    @{ int *_int_p = 0L; *_int_p = 1; @}
@end example

@c --------------------------------

@node Extensions, Error Codes, Dumping Core, Programming
@section Additional Non-standard Routines

@cindex extensions

The library has a number of variables that are not a standard part of
most malloc libraries:

@c --------------------------------

@cindex dmalloc_errno number
@cindex internal error number
@cindex error number

@deftypevar int dmalloc_errno

This variable stores the internal dmalloc library error number like errno
does for the system calls.  It can be passed to @code{dmalloc_strerror()}
(see below) to get a string version of the error.  It will have a value
of zero if the library has not detected any problems.

@end deftypevar

@c --------------------------------

@cindex dmalloc_logpath variable
@cindex logfile name

@deftypevar char* dmalloc_logpath

This variable can be used to set the dmalloc log filename.  The env
variable @code{DMALLOC_LOGFILE} overrides this variable.

@c --------------------------------

@end deftypevar

Additionally the library provides a number of non-standard malloc
routines:

@c --------------------------------

@cindex dmalloc_shutdown function
@cindex shutdown the library

@deftypefun void dmalloc_shutdown ( void )

@cindex atexit function
@cindex on_exit function

This function shuts the library down and logs the final statistics and
information especially the non-freed memory pointers.  The library has
code to support auto-shutdown if your system has the @code{on_exit()}
call, @code{atexit()} call, or compiler destructor support (see
@file{conf.h}).  If you do not have these, then @code{dmalloc_shutdown}
should be called right before @code{exit()} or as the last function in
@code{main()}.

@example
main()
@{
        @dots{}
        dmalloc_shutdown();
        exit(0);
@}
@end example
@end deftypefun

@c --------------------------------

@cindex dmalloc_verify function
@cindex verify pointers
@cindex verify the heap

@deftypefun int dmalloc_verify ( char * @var{pnt} )

This function verifies individual memory pointers that are suspect of
memory problems.  To check the entire heap pass in a NULL or 0 pointer.
The routine returns DMALLOC_VERIFY_ERROR or DMALLOC_VERIFY_NOERROR.

@emph{NOTE}: @samp{dmalloc_verify()} can only check the heap with the
functions that have been enabled.  For example, if fence-post checking
is not enabled, @samp{dmalloc_verify()} cannot check the fence-post
areas in the heap.
@end deftypefun

@c --------------------------------

@cindex dmalloc_debug function
@cindex override debug settings
@cindex set debug functionality flags

@deftypefun unsigned-int dmalloc_debug ( const unsigned int @var{flags} )

This routine sets the debug functionality flags and returns the
previous flag value.  It is helpful in server or cgi-bin programs
where environmental variables cannot be used.  @xref{Debugging A
Server}.  For instance, if debugging should never be enabled for a
program, a call to @code{dmalloc_debug(0)} as the first call in
@code{main()} will disable all the memory debugging from that point
on.

@emph{NOTE}: you cannot add or remove certain flags such as signal
handlers since they are setup at initialization time only.

@emph{NOTE}: you can also use @code{dmalloc_debug_setup} below.

@end deftypefun

@c --------------------------------

@cindex dmalloc_debug_current function
@cindex current debug value

@deftypefun unsigned-int dmalloc_debug_current ( void )

This routine returns the current debug functionality value value.  This
allows you to save a copy of the debug dmalloc settings to be changed
and then restored later.

@end deftypefun

@c --------------------------------

@cindex dmalloc_debug_setup
@cindex override debug settings
@cindex set debug functionality flags
@cindex setup debug flags
@cindex string of debug tokens

@deftypefun void dmalloc_debug_setup ( const char * @var{options_str} )

This routine sets the global debugging functionality as an option
string.  Normally this would be passed in in the DMALLOC_OPTIONS
environmental variable.  This is here to override the env or for
circumstances where modifying the environment is not possible or does
not apply such as servers or cgi-bin programs.  @xref{Debugging A
Server}.

Some examples:

@example
/*
 * debug tokens high, threaded lock-on at 20,
 * log to dmalloc.%p (pid)
 */
dmalloc_debug_setup("debug=0x4f46d03,lockon=20,log=dmalloc.%p");

/*
 * turn on some debug tokens directly and log to the
 * file 'logfile'
 */
dmalloc_debug_setup(
  "log-stats,log-non-free,check-fence,log=logfile");
@end example

@end deftypefun

@c --------------------------------

@cindex dmalloc_examine function
@cindex examine a pointer
@cindex pointer information

@deftypefun int dmalloc_examine ( const DMALLOC_PNT @var{pnt}, DMALLOC_SIZE * @var{user_size_p}, DMALLOC_SIZE * @var{total_size_p}, char ** @var{file_p}, int * @var{line_p}, DMALLOC_PNT * @var{ret_addr_p}, unsigned long * @var{user_mark_p}, unsigned long * @var{seen_p} )

@cindex seen count
This function returns the size of a pointer's allocation as well as the
total size given including administrative overhead, file and line or the
return-address from where it was allocated, the last pointer when the
pointer was "used", and the number of times the pointer has been "seen".
It will return DMALLOC_NOERROR or DMALLOC_ERROR depending on whether pnt
is good or not.

@emph{NOTE}: This function is @emph{certainly} not provided by most if
not all other malloc libraries.

@end deftypefun

@c --------------------------------

@cindex dmalloc_track function
@cindex track memory calls

@deftypefun void dmalloc_track ( const dmalloc_track_t @var{track_func} )

Register an allocation tracking function which will be called each time
an allocation occurs.  Pass in NULL to disable.  To take a look at what
information is provided, see the dmalloc_track_t function typedef in
dmalloc.h.

@end deftypefun

@c --------------------------------

@cindex dmalloc_mark function
@cindex memory position marker
@cindex mark memory position
@cindex LOG_ITERATION
@cindex interaction count
@cindex transaction count

@deftypefun unsigned-long dmalloc_mark ( void )

Return to the caller the current "mark" which can be used later to log
the pointers which have changed since this mark with the
@code{dmalloc_log_changed} function.  Multiple marks can be saved and
used.

This is very useful when using the library with a server which does
not exit.  You can then save a mark before a transaction or event
happens and then check to see what has changed using the
@code{dmalloc_log_changed} function below.  @xref{Debugging A Server}.

If you @code{LOG_ITERATION} enabled in your @file{settings.h} file then
the entries in the log file will be prepended with the number of memory
transactions that the library has handled so far.  You can also enable
@code{LOG_PNT_ITERATION} in @file{settings.h} to store the memory
transaction number with each pointer.

@end deftypefun

@c --------------------------------

@cindex dmalloc_memory_allocated function
@cindex memory allocated function
@cindex number of bytes allocated

@deftypefun unsigned-long dmalloc_memory_allocated ( void )

Return to the caller the total number of bytes that have been allocated
by the library.  This is not the current in use but the total number of
bytes returned by allocation functions.

@end deftypefun

@c --------------------------------

@cindex page size
@cindex size of memory pages

@deftypefun unsigned-int dmalloc_page_size ( void )

Return to the caller the memory page-size being used by the library.
This should be the same value as the one returned by the
@code{getpagesize()} function, if available.

@end deftypefun

@c --------------------------------

@cindex count changed
@cindex count number of bytes changed since mark
@cindex number bytes changed since mark

@deftypefun unsigned-long dmalloc_count_changed ( const unsigned long @var{mark}, const int @var{not_freed_b}, const int @var{free_b} )

Count the pointers that have changed since the mark which was returned by
@code{dmalloc_mark}.  If @code{not_freed_b} is set to non-0 then count the
pointers that have not been freed.  If @code{free_b} is set to non-0
then count the pointers that have been freed.

This can be used in conjunction with the @code{dmalloc_mark()}
function to help servers which never exit ensure that transactions or
events are not leaking memory.  @xref{Debugging A Server}.

@example
unsigned long mark = dmalloc_mark() ;
@dots{}
assert(dmalloc_count_changed(mark, 1, 0) == 0) ;
@end example

@end deftypefun

@c --------------------------------

@cindex dmalloc_log_stats function
@cindex log statistics
@cindex statistics logging

@deftypefun void dmalloc_log_stats ( void )

This routine outputs the current dmalloc statistics to the log file.
@end deftypefun

@c --------------------------------

@cindex dmalloc_log_unfreed function
@cindex log unfreed memory
@cindex unfreed memory log

@deftypefun void dmalloc_log_unfreed ( void )

This function logs the unfreed-memory information to the log file.
This is also useful to log the currently allocated points to the log
file to be compared against another dump later on.

@end deftypefun

@c --------------------------------

@cindex dmalloc_log_changed function
@cindex changed memory log
@cindex checkpoint memory usage
@cindex log memory changes

@deftypefun void dmalloc_log_changed ( const unsigned long @var{mark}, const int @var{not_freed_b}, const int @var{freed_b}, const int @var{details_b} )

Log the pointers that have changed since the mark which was returned by
@code{dmalloc_mark}.  If @code{not_freed_b} is set to non-0 then log the
pointers that have not been freed.  If @code{free_b} is set to non-0
then log the pointers that have been freed.  If @code{details_b} set to
non-0 then log the individual pointers that have changed otherwise just
log the summaries.

This can be used in conjunction with the @code{dmalloc_mark()}
function to help servers which never exit find transactions or events
which are leaking memory.  @xref{Debugging A Server}.

@end deftypefun

@c --------------------------------

@cindex dmalloc_vmessage function
@cindex write message to logfile
@cindex logfile message writer

@deftypefun void dmalloc_vmessage ( const char * @var{format}, va_list @var{args} )

Write a message into the dmalloc logfile using vprintf-like arguments.

@end deftypefun

@c --------------------------------

@cindex dmalloc_message function
@cindex write message to logfile
@cindex logfile message writer

@deftypefun void dmalloc_message ( const char * @var{format}, @var{...} )

Write a message into the dmalloc logfile using printf-like arguments.

@end deftypefun

@c --------------------------------

@cindex dmalloc_get_stats function
@cindex get heap statistics
@cindex heap statistics function
@cindex report heap statistics

@deftypefun void dmalloc_get_stats ( DMALLOC_PNT * @var{heap_low_p}, DMALLOC_PNT * @var{heap_high_p}, unsigned long * @var{total_space_p}, unsigned long * @var{user_space_p}, unsigned long * @var{current_allocated_p}, unsigned long * @var{current_pnt_np}, unsigned long * @var{max_allocated_p}, unsigned long * @var{max_pnt_np}, unsigned long * @var{max_one_p})

This function return a number of statistics about the current heap.
The pointers @code{heap_low_p} and @code{heap_high_p} will be set to
the low and high spots in the heap.  @code{total_space_p} will be set
to the total space in the heap including user space, administrative
space, and overhead.  @code{user_space_p} will be set to the space
given to the user process (allocated and free space).
@code{current_allocated_p} will be set to the current allocated space
given to the user process.  @code{current_pnt_np} will be set to the
current number of pointers allocated by the user process.
@code{max_allocated_p} will be set to the maximum allocated space
given to the user process.  @code{max_pnt_np} will be set to the
maximum number of pointers allocated by the user process.
@code{max_on_p} will be set to the maximum space allocated with one
call by the user process.

@end deftypefun

@c --------------------------------

@cindex dmalloc_strerror function
@cindex string error message
@cindex error message

@deftypefun const-char* dmalloc_strerror ( const int @var{error_number} )

This function returns the string representation of the error value in
@code{error_number} (which probably should be dmalloc_errno).  This
allows the logging of more verbose memory error messages.

You can also display the string representation of an error value by a
call to the @file{dmalloc} program with a @samp{-e #} option.
@xref{Dmalloc Program}.

@end deftypefun

@c --------------------------------

@node Error Codes, Disabling the Library, Extensions, Programming
@section Description of the Internal Error Codes

@cindex error codes

The following error codes are defined in @file{error_val.h}.  They are
used by the library to indicate a detected problem.  They can be caused
by the user (@samp{ERROR_TOO_BIG}) or can indicate an internal library
problem (@samp{ERROR_SLOT_CORRUPT}).  The @file{dmalloc} utility can
give you the string version of the error with the @code{-e} argument:

@example
$ dmalloc -e 60
dmalloc: dmalloc_errno value '60' = 
   'pointer is not on block boundary'
@end example

Here are the error codes set by the library.  They are non contiguous on
purpose because I add and delete codes all of the time and there are
sections for various error-code types.

@table @code

@c --------------------------------

@cindex 1, error code
@cindex error code, 1
@cindex no error
@cindex ERROR_NONE

@item 1 (ERROR_NONE) no error
No error.  It is good coding practice to set the no-error code to be
non-0 value because it forces you to set it explicitly.

@c --------------------------------

@cindex 2, error code
@cindex error code, 2
@cindex invalid errno value
@cindex errno value is not valid
@cindex INVALID_ERROR

@item 2 (INVALID_ERROR)
Invalid error number.  If the library outputs this error then your
dmalloc utility may be out of date with the library you linked
against.  This will be returned with all error codes not listed here.

@c --------------------------------

@cindex 10, error code
@cindex error code 10
@cindex initialization and setup failed
@cindex bad setup error
@cindex ERROR_BAD_SETUP

@item 10 (ERROR_BAD_SETUP) initialization and setup failed
Bad setup value.  This is currently unused but it is intended to report
on invalid setup configuration information.

@c --------------------------------

@cindex lock-on
@cindex recursion
@cindex in twice error
@cindex 11, error code
@cindex error code 11
@cindex malloc library has gone recursive
@cindex ERROR_IN_TWICE

@item 11 (ERROR_IN_TWICE) malloc library has gone recursive
Library went recursive.  This usually indicates that you are not using
the threaded version of the library.  Or if you are then you are not
using the @samp{-o} "lock-on" option.  @xref{Using With Threads}.

@c --------------------------------

@cindex lock-on not configured
@cindex 13, error code
@cindex error code 13
@cindex thread locking has not been configured
@cindex ERROR_LOCK_NOT_CONFIG

@item 13 (ERROR_LOCK_NOT_CONFIG) thread locking has not been configured
Thread locking has not been configured.  This indicates that you
attempted to use the @samp{-o} "lock-on" option without linking with
the thread version of the library.  You should probably be using
@code{-ldmallocth} @emph{not} @code{-ldmalloc} when you are linking.
Or you should include @code{@dots{}/lib/libdmallocth.a} on your
compilation line.

@c --------------------------------

@cindex error-free-null token
@cindex free null token
@cindex 20, error code
@cindex error code 20
@cindex pointer is null error
@cindex is null error
@cindex ERROR_IS_NULL

@item 20 (ERROR_IS_NULL) pointer is null
Pointer is null.  The program passed a NULL (0L) pointer to @code{free}
and you have the @code{error-free-null} token enabled.

@c --------------------------------

@cindex pointer not in heap error
@cindex not in heap error
@cindex 21, error code
@cindex error code 21
@cindex pointer is not pointing to heap data space
@cindex ERROR_NOT_IN_HEAP

@item 21 (ERROR_NOT_IN_HEAP) pointer is not pointing to heap data space
Pointer is not pointing to heap data space.  This means that the program
passed an out-of-bounds pointer to @code{free} or @code{realloc}.  This
could be someone trying to work with a wild pointer or trying to free a
pointer from a different source than @code{malloc}.

@c --------------------------------

@cindex not found error
@cindex pointer not found error
@cindex 22, error code
@cindex error code 22
@cindex cannot locate pointer in heap
@cindex ERROR_NOT_FOUND
@cindex mmap
@cindex sbrk

@item 22 (ERROR_NOT_FOUND) cannot locate pointer in heap
Cannot locate pointer in heap.  The user passed in a pointer which the
heap did not know about.  Either this pointer was allocated by some
other mechanism (like @code{mmap} or @code{sbrk} directly) or it is a
random invalid pointer.

In some rare circumstances, sometimes seen with shared libraries, there
can be two separate copies of the dmalloc library in a program.  Each
one does not know about the pointers allocated by the other.

@c --------------------------------

@cindex 23, error code
@cindex error code 23
@cindex found pointer the user was looking for
@cindex is found error
@cindex ERROR_IS_FOUND

@item 23 (ERROR_IS_FOUND) found pointer the user was looking for
This indicates that the pointer specified in the address part of the
environmental variable was discovered by the library.  @xref{Environment
Variable}.  This error is useful so you can put a breakpoint in a
debugger to find where a particular address was allocated.  @xref{Using
With a Debugger}.

@c --------------------------------

@cindex 24, error code
@cindex error code 24
@cindex possibly bad .c filename pointer
@cindex bad file error
@cindex ERROR_BAD_FILE

@item 24 (ERROR_BAD_FILE) possibly bad .c filename pointer
A possibly invalid filename was discovered in the dmalloc administrative
sections.  This could indicate some corruption of the internal tables.
It also could mean that you have a source file whose name is longer than
100 characters.  See @code{MAX_FILE_LENGTH} in the @file{settings.dist}
file.

@c --------------------------------

@cindex 25, error code
@cindex error code 25
@cindex possibly bad .c file line-number
@cindex bad line error
@cindex ERROR_BAD_LINE

@item 25 (ERROR_BAD_LINE) possibly bad .c file line-number
A line-number was out-of-bounds in the dmalloc administrative sections.
This could indicate some corruption of the internal tables.  It also
could mean that you have a source file containing more than @code{30000}
lines of code.  See @code{MAX_LINE_NUMBER} in the @file{settings.dist}
file.

@c --------------------------------

@cindex 26, error code
@cindex error code 26
@cindex failed under picket-fence magic-number check
@cindex under fence error
@cindex ERROR_UNDER_FENCE

@item 26 (ERROR_UNDER_FENCE) failed UNDER picket-fence magic-number check
This indicates that a pointer had its lower bound picket-fence magic
number overwritten.  If the @code{check-fence} token is enabled, the
library writes magic values above and below allocations to protect
against overflow.  Most likely this is because a pointer below it went
past its allocate and wrote into the next pointer's space.

@c --------------------------------

@cindex 27, error code
@cindex error code 27
@cindex failed over picket-fence magic-number check
@cindex over fence error
@cindex ERROR_OVER_FENCE

@item 27 (ERROR_OVER_FENCE) failed OVER picket-fence magic-number check
This indicates that a pointer had its upper bound picket-fence magic
space overwritten.  If the @code{check-fence} token is enabled, the
library writes magic values above and below allocations to protect
against overflow.  Most likely this is because an array or string
allocation wrote past the end of the allocation.

Check for improper usage of @code{strcat}, @code{sprintf},
@code{strcpy}, and any other functions which work with strings and do
not protect themselves by tracking the size of the string.  These
functions should @emph{always} be replaced with: @code{strncat},
@code{snprintf}, @code{strncpy}, and others.

@c --------------------------------

@cindex 28, error code
@cindex error code 28
@cindex use of pointer would exceed allocation
@cindex would overwrite error
@cindex ERROR_WOULD_OVERWRITE

@item 28 (ERROR_WOULD_OVERWRITE) use of pointer would exceed allocation
This error is generated by the function pointer checking code usually
enabled with the @code{check-funcs} token.  Dmalloc overloads a number of
string and memory copying functions and verifies that the buffers (if
allocated in the heap) would not be overwritten by the function.

@c --------------------------------

@cindex 30, error code
@cindex error code 30
@cindex pointer is not to start of memory block
@cindex not start block error
@cindex ERROR_NOT_START_BLOCK

@item 30 (ERROR_NOT_START_BLOCK) pointer is not to start of memory block
This indicates that the user passed in a pointer to be freed or
reallocated that was not at the start of the allocation.  You would get
this error, for example, if you allocate and get pointer @code{X} but
then try to free @code{X+1}.

@c --------------------------------

@cindex 40, error code
@cindex error code 40
@cindex invalid allocation size
@cindex bad size error
@cindex ERROR_BAD_SIZE

@item 40 (ERROR_BAD_SIZE) invalid allocation size
This error indicates that a size value in the internal structures of the
library were corrupted.  This could be a random pointer problem, pointer
overflow, or some other corruption.

@c --------------------------------

@cindex 41, error code
@cindex error code 41
@cindex largest maximum allocation size exceeded
@cindex too big error
@cindex ERROR_TOO_BIG

@item 41 (ERROR_TOO_BIG) largest maximum allocation size exceeded
An allocation asked for memory larger than the configured maximum.  This
is a user configured setting.  See @code{LARGEST_ALLOCATION} in the
@file{settings.dist} file.  It is used to protect against wild
allocation sizes.  If you have super large allocation sizes then you
should tune the @code{LARGEST_ALLOCATION} value appropriately.

@c --------------------------------

@cindex 43, error code
@cindex error code 43
@cindex could not grow heap by allocating memory
@cindex alloc failed error
@cindex ERROR_ALLOC_FAILED

@item 43 (ERROR_ALLOC_FAILED) could not grow heap by allocating memory
The library could not allocate more heap space and the program has run
out of memory.  This could indicate that you've overflowed some system
imposed limit.  On many operation systems, the @code{ulimit} call can
tune system defaults.  The library uses a lot more memory compared to
the system malloc library because it stores a lot more information
about the allocated pointers.

@cindex mmap
@cindex USE_MMAP conf.h option

@emph{NOTE}: This also may be due to an inability of your operating
system to use the @code{mmap} system call to allocate memory.  You may
need to force the @code{USE_MMAP} setting to be 0.  Please use the
forums at URL @uref{http://dmalloc.com/} to report issues with this.

@c --------------------------------

@cindex 45, error code
@cindex error code 45
@cindex memory limit
@cindex over user specified allocation limit error
@cindex over limit error
@cindex ERROR_OVER_LIMIT

@item 45 (ERROR_OVER_LIMIT) over user specified allocation limit
The library has allocated more memory than was specified in the
memory-limit environmental variable.  @xref{Environment Variable}.

@c --------------------------------

@cindex 60, error code
@cindex error code 60
@cindex pointer is not on block boundary
@cindex not on block boundary error
@cindex ERROR_NOT_ON_BLOCK

@item 60 (ERROR_NOT_ON_BLOCK) pointer is not on block boundary
The user tried to free or realloc a pointer that was not pointing to a
block boundary.  You would get this error, for example, if you allocate
and get pointer @code{X} but then try to free @code{X+1}.

@c --------------------------------

@cindex 61, error code
@cindex error code 61
@cindex tried to free previously freed pointer
@cindex already free error
@cindex ERROR_ALREADY_FREE

@item 61 (ERROR_ALREADY_FREE) tried to free previously freed pointer
The user tried to free a pointer than has already been freed.  This is a
very common mistake and can lead to serious problems.  It can be because
a destructor is being called twice for some reason.  Although tracking
down the specific source is highly recommended, it is good to set
pointers to NULL (0L) after you free them as a rule.

@c --------------------------------

@cindex 67, error code
@cindex error code 67
@cindex free space has been overwritten
@cindex free overwritten error
@cindex ERROR_FREE_OVERWRITTEN
@cindex free-blank
@cindex check-blank

@item 67 (ERROR_FREE_OVERWRITTEN) free space has been overwritten
If either the @code{free-blank} or @code{check-blank} tokens are enabled
then the library will overwrite memory when it is freed with the
``dmalloc-free'' byte (hex 0xdf, octal 0337, decimal 223).  If the
program writes into this space, then the library will detect the write
and trigger this error.  This could indicate that the program is using a
pointer after it has been freed.

@c --------------------------------

@cindex 70, error code
@cindex error code 70
@cindex bad admin structure list
@cindex admin list error
@cindex ERROR_ADMIN_LIST

@item 70 (ERROR_ADMIN_LIST) bad admin structure list
An internal corruption in the library's administrative structures has
been detected.  This could be a random pointer problem, pointer
overflow, or some other corruption.

@c --------------------------------

@cindex 72, error code
@cindex error code 72
@cindex internal address list corruption
@cindex address list error
@cindex ERROR_ADDRESS_LIST

@item 72 (ERROR_ADDRESS_LIST) internal address list corruption
An internal corruption in the library's administrative structures has
been detected.  This could be a random pointer problem, pointer
overflow, or some other corruption.

@c --------------------------------

@cindex 73, error code
@cindex error code 73
@cindex internal memory slot corruption
@cindex slot corrupt error
@cindex ERROR_SLOT_CORRUPT

@item 73 (ERROR_SLOT_CORRUPT) internal memory slot corruption
An internal corruption in the library's administrative structures has
been detected.  This could be a random pointer problem, pointer
overflow, or some other corruption.

@end table

@c --------------------------------

@node Disabling the Library, Using With C++, Error Codes, Programming
@section How to Disable the library

@cindex disabling library checking
@cindex disabling the library
@cindex dmalloc_debug
@cindex restoring library flags
@cindex turning off library flags

If you would like to disable the library's detailed checking features
during a particularly allocation intensive section of code, you can do
something like the following:

@example
        unsigned int dmalloc_flags;
        @dots{}
        /* turn off all debug flags and save a copy of old value */
        dmalloc_flags = dmalloc_debug(0);

        /* section of a lot of allocations */
        @dots{}
        /* end of section */

        /* restore the dmalloc flag setting */
        dmalloc_debug(dmalloc_flags);
@end example

When you are finished with the development and debugging sessions, you
may want to disable the dmalloc library and put in its place either the
system's memory-allocation routines, gnu-malloc, or maybe your own.
Attempts have been made to make this a reasonably painless process.  The
ease of the extraction depends heavily on how many of the library's
features your made use of during your coding.

Reasonable suggestions are welcome as to how to improve this process
while maintaining the effectiveness of the debugging.

@itemize @bullet
@item
If you want to @emph{totally} disable the dmalloc library then you will
need to recompile all the C files that include @file{dmalloc.h} while
defining @code{DMALLOC_DISABLE}.  This will cause the dmalloc macros to
not be applied.  @xref{Allocation Macros}.

@example
cc -g -DDMALLOC_DISABLE file.c
@end example

An alternative is to surround the @code{dmalloc.h} inclusion or any
direct dmalloc references with an @code{#ifdef DMALLOC} and then just
remove the -DDMALLOC.

@example
#ifdef DMALLOC
#include "dmalloc.h"
#endif

main()
@{
  @dots{}

#ifdef DMALLOC
  dmalloc_verify(0L);
#endif
  return 0;
@}
@end example

@example
// to get dmalloc information
$ cc -DDMALLOC main.c

// without dmalloc information
$ cc main.c
@end example

@item
If you compiled any of your source modules with @code{DMALLOC_FUNC_CHECK}
defined then you must first recompile all those modules without the flag
enabled.

If you have disabled dmalloc with the @code{DMALLOC_DISABLED} flag or
never included @file{dmalloc.h} in any of your C files, then you will
not need to recompile your sources when you need to disable the library.

If you get unresolved references like @code{_dmalloc_malloc} or
@code{_dmalloc_bcopy} then something was not disabled as it should have
been.
@end itemize

@c --------------------------------

@node Using With C++, Using With a Debugger, Disabling the Library, Programming
@section Using the Library with C++

@cindex c++ usage
@cindex dmallocc.cc file

@cindex libdmallocxx.a
For those people using the C++ language, the library tries to configure
and build @file{libdmallocxx.a} library.  This library should be linked
into your C++ programs instead of @file{libdmalloc.a}.

@cindex dmallocc.cc
Dmalloc is not as good with C++ as C because the dynamic memory routines
in C++ are @code{new()} and @code{delete()} as opposed to
@code{malloc()} and @code{free()}.  Since new and delete are usually not
used as functions but rather as @code{x = new type}, there is no easy
way for dmalloc to pass in file and line information unfortunately.  The
@file{libdmallocxx.a} library provides the file @file{dmallocc.cc} which
effectively redirects @code{new} to the more familiar @code{malloc} and
@code{delete} to the more familiar @code{free}.

@emph{NOTE}: The author is not a C++ hacker so feedback in the form of
other hints and ideas for C++ users would be much appreciated.

@c --------------------------------

@node Using With a Debugger, Using With Threads, Using With C++, Programming
@section Using Dmalloc With a Debugger

@cindex debugger usage with dmalloc
@cindex using a debugger with dmalloc

Here are a number of possible scenarios for using the dmalloc library to
track down problems with your program.

You should first enable a logfile filename and turn on a set of debug
features.  You can use @kbd{dmalloc -l logfile low} to accomplish this.
If you are interested in having the error messages printed to your
terminal as well, enable the @code{print-messages} token by typing
@kbd{dmalloc -p print-messages} afterwards.  @xref{Dmalloc Program}.

@cindex gdb
@cindex dmalloc_error() routine
Now you can enter your debugger (I use the @emph{excellent} GNU debugger
gdb), and put a break-point in @code{dmalloc_error()} which is the
internal error routine for the library.  When your program is run, it
will stop there if a memory problem is detected.

If you are using GDB, I would recommend adding the contents of
@file{dmalloc.gdb} in the @file{contrib} subdirectory to your
@file{.gdbinit} file in your home directory.  This enables the
@code{dmalloc} command which will prompt you for the arguments to the
dmalloc command and will set a break point in @code{dmalloc_error()}
automatically.

@cindex shared libraries with gdb
@cindex gdb with shared libraries
If you are using shared libraries, you may want to execute the following
commands initially to load in dmalloc and other library symbols:

@example
(gdb) sharedlibrary
(gdb) add-shared-symbol-files
@end example

@menu
* General Errors::                Diagnosing general problems with a debugger.
* Memory Leaks::                  Tracking down non-freed memory.
* Fence-Post Overruns::           Diagnosing fence-post overwritten memory.
* Translate Return Addresses::    Convert ra return-addresses into a location.
@end menu

@c --------------------------------

@node General Errors, Memory Leaks, Using With a Debugger, Using With a Debugger
@subsection Diagnosing General Problems with a Debugger

@cindex diagnosing errors
@cindex general errors
@cindex mmap
@cindex sbrk

If your program stops at the @code{dmalloc_error()} routine then one
of a number of problems could be happening.  Incorrect arguments could
have been passed to a malloc call: asking for negative number of
bytes, trying to realloc a non-heap pointer, etc..  There also could
be a problem with the system's allocations: you've run out of memory,
some other function in your program is using the heap allocation
functions @code{mmap} or @code{sbrk}, etc..  However, it is most
likely that some code that has been executed was naughty.

To get more information about the problem, first print via the debugger
the dmalloc_errno variable to get the library's internal error code.
You can suspend your debugger and run @kbd{dmalloc -e
value-returned-from-print} to get an English translation of the error.
A number of the error messages are designed to indicate specific
problems with the library administrative structures and may not be
user-friendly.

If the problem was due to the arguments or system allocations then the
source of the problem has been found.  However, if some code did
something wrong, you may have some more work to do to locate the actual
problem.  The @code{check-heap} token should be enabled and the interval
setting disabled or set to a low value so that the library can find the
problem as close as possible to its source.  The code that was execute
right before the library halted, can then be examined closely for
irregularities.  @xref{Debug Tokens}, @xref{Dmalloc Program}.

@cindex dmalloc_verify() routine

You may also want to put calls to @code{dmalloc_verify(0)} in your code
before the section which generated the error.  This should locate the
problem faster by checking the library's structures at that point.
@xref{Extensions}.

@c --------------------------------

@node Memory Leaks, Fence-Post Overruns, General Errors, Using With a Debugger
@subsection Tracking Down Non-Freed Memory

@cindex memory leaks
@cindex leaking memory

So you've run your program, examined the log-file and discovered (to
your horror) some un-freed memory.  Memory leaks can become large
problems since even the smallest and most insignificant leak can starve
the program given the right circumstances.

@example
not freed: '0x45008' (12 bytes) from 'ra=0x1f8f4'
not freed: '0x45028' (12 bytes) from 'unknown'
not freed: '0x45048' (10 bytes) from 'argv.c:1077'
  known memory not freed: 1 pointer, 10 bytes
unknown memory not freed: 2 pointers, 24 bytes
@end example

Above you will see a sample of some non-freed memory messages from the
logfile.  In the first line the @samp{0x45008} is the pointer that was
not freed, the @samp{12 bytes} is the size of the unfreed block, and the
@samp{ra=0x1f8f4} or return-address shows where the allocation
originated from.  @xref{Translate Return Addresses}.

The systems which cannot provide return-address information show
@samp{unknown} instead, as in the 2nd line in the sample above.

The @samp{argv.c:1077} information from the 3rd line shows the file and
line number which allocated the memory which was not freed.  This
information comes from the calls from C files which included
@file{dmalloc.h}.  @xref{Allocation Macros}.

At the bottom of the sample it totals the memory for you and breaks it
down to known memory (those calls which supplied the file/line
information) and unknown (the rest).

Often, you may allocate memory in via @code{strdup()} or another
routine, so the logfile listing where in the @code{strdup} routine the
memory was allocated does not help locate the true source of the memory
leak -- the routine that called @code{strdup}.  Without a mechanism to
trace the calling stack, there is no way for the library to see who the
caller of the caller (so to speak) was.

@cindex STORE_SEEN_COUNT conf.h option
@cindex pointer seen count
@cindex seen count

However, there is a way to track down unfreed memory in this
circumstance.  You need to compile the library with
@code{STORE_SEEN_COUNT} defined in @file{conf.h}.  The library will then
record how many times a pointer has been allocated or freed.  It will
display the unfreed memory as:

@example
not freed: '0x45008|s3' (12 bytes) from 'ra=0x1f8f4'
@end example

The @code{STORE_SEEN_COUNT} option adds a @samp{|s#} qualifier to the
address.  This means that the address in question was seen @samp{#} many
times.  In the above example, the address @samp{0x45008} was seen
@samp{3} times.  The last time it was allocated, it was not freed.

How can a pointer be ``seen'' 3 times?  Let say you @code{strdup} a
string of 12 characters and get address @samp{0x45008} -- this is #1
time the pointer is seen.  You then free the pointer (seen #2) but later
@code{strdup} another 12 character string and it gets the @samp{0x45008}
address from the free list (seen #3).

So to find out who is allocating this particular 12 bytes the 3rd time,
try @kbd{dmalloc -a 0x45008:3}.  The library will stop the program the
third time it sees the @samp{0x45008} address.  You then enter a
debugger and put a break point at @code{dmalloc_error}.  Run the program
and when the breakpoint is reached you can examine the stack frame to
determine who called @code{strdup} to allocate the pointer.

To not bother with the @code{STORE_SEEN_COUNT} feature, you can also run
your program with the @code{never-reuse} token enabled.  This token will
cause the library to never reuse memory that has been freed.  Unique
addresses are always generated.  This should be used with caution since
it may cause your program to run out of memory.

@c --------------------------------

@node Fence-Post Overruns, Translate Return Addresses, Memory Leaks, Using With a Debugger
@subsection Diagnosing Fence-Post Overwritten Memory

@cindex fence-post errors

For a definition of fence-posts please see the ``Features'' section.
@xref{Features}.

To detect fence-post overruns, you need to enable the @samp{check-fence}
token.  @xref{Debug Tokens}.  This pads your allocations with some extra
bytes at the front and the end and watches the space to make sure that
they don't get overwritten.  @emph{NOTE:} The library cannot detect if
this space gets read, only written.

If you have encountered a fence-post memory error, the logfile should be
able to tell you the offending address.

@example
free: failed UNDER picket-fence magic-number checking: 
pointer '0x1d008' from 'dmalloc_t.c:427'
Dump of proper fence-bottom bytes: '\e\253\300\300\e\253\300\300'
Dump of '0x1d008'-8: '\e\253\300\300WOW!\003\001pforger\023\001\123'
@end example

The above sample shows that the pointer @samp{0x1d008} has had its lower
fence-post area overwritten.  This means that the code wrote below the
bottom of the address or above the address right below this one.  In the
sample, the string that did it was @samp{WOW!}.

The library first shows you what the proper fence-post information
should look like, and then shows what the pointer's bad information was.
If it cannot print the character, it will display the value as
@samp{\ddd} where ddd are three octal digits.

By enabling the @code{check-heap} debugging token and assigning the
interval setting to a low number, you should be able to locate
approximately when this problem happened.  @xref{Debug Tokens},
@xref{Dmalloc Program}.

@c --------------------------------

@node Translate Return Addresses,, Fence-Post Overruns, Using With a Debugger
@subsection Translating Return Addresses into Code Locations

@cindex ra_info.pl

The following gdb commands help you translate the return-addresses (ra=)
entries in the logfile into locations in your code.  I've provided a
@file{ra_info.pl} perl script in the @file{contrib/} directory with the
dmalloc sources which seems to work well with gdb.  But, if you need to
do it manually, here are the commands in gdb to use.

@cindex return-address translation
@cindex caller address translation

@example
# you may need to add the following commands to load in shared libraries
(gdb) sharedlibrary
(gdb) add-shared-symbol-files

(gdb) x 0x10234d
0x10234d <_findbuf+132>: 0x7fffceb7

(gdb) info line *(0x82cc)
Line 1092 of argv.c starts at pc 0x7540 and ends at 0x7550.
@end example

In the above example, gdb was used to find that the two non-freed memory
pointers were allocated in @code{_findbuf()} and in file argv.c line
1092 respectively.  The @samp{x address} (for examine) can always be
used on the return-addresses but the @samp{info line *(address)} will
only work if that file was compiled using the @kbd{-g} option and has
not been stripped.  This limitation may not be true in later versions of
gdb.

@c --------------------------------

@node Using With Threads, Using With Cygwin, Using With a Debugger, Programming
@section Using the Library with a Thread Package

@cindex threads
@cindex pthreads

Threads are special operating system facilities which allow your
programs to have multiple threads of execution (hence the name).  In
effect your program can be doing a number of things ``at the same
time''.  This allows you to take full advantage of modern operating
system scheduling and multi-processor hardware.  If I've already lost
you or if any of the terminology below does not make sense, see manuals
about POSIX threads (pthreads) before going any further.  O'Reilly
publishes a pretty good pthreads manual for example.

To use dmalloc with your threaded program, you will first need to make
sure that you are linking with @file{libdmallocth.a} which is the
threaded version of the library.  The support for threads in dmalloc
should be adequate for most if not all testing scenarios.  It provides
support for mutex locking itself to protect against race conditions that
result in multiple simultaneous execution.  One of the major problems is
that most thread libraries uses malloc themselves.  Since all of
dmalloc's initialization happens when a call to malloc is made, we may
be attempting to initialize or lock the mutex while the thread library
is booting up.  A very bad thing since thread libraries don't expect to
recurse.

@cindex lock on

The solution to this problem is to have the library not initialize or
lock its mutex variable until after a certain number of allocation calls
have been completed.  If the library does not wait before initializing
the locks, the thread library will probably core dump.  If it waits too
long then it can't protect itself from multiple execution and it will
abort or other bad things might happen.  You adjust the number of times
to wait at runtime with the @samp{lock-on} option to the dmalloc program
(for example @kbd{dmalloc -o 20}).  @xref{Dmalloc Program}.  Times
values between 5 and 30 are probably good although operating systems
will vary significantly.  You know its too low if your program
immediately core dumps and too high if the dmalloc library says its gone
recursive although with low values, you might get either problem.

An additional complexity is when we are initializing the lock before
mutex locking around the library.  As mentioned, the initialization
itself may generate a malloc call causing the library to go recursive
and the pthread library to possibly core dump. With the THREAD_INIT_LOCK
setting defined in @file{settings.h}, you can tune how many times before
we start locking to try and initialize the mutex lock.  It defaults to 2
which seems to work for me.  If people need to have this runtime
configurable or would like to present an alternative default, please let
me know.

So to use dmalloc with a threaded program, follow the following steps
carefully.

@enumerate

@item Follow the installation instructions on how to configure,
make, and install the library but make sure to add the
@kbd{--enable-threads} argument to configure.  @xref{Installation}.

@item Typing @kbd{make} should be enough to build the threaded versions
of the libraries including @file{libdmallocth.a}.

@item Link the dmalloc threaded library into your program.  The dmalloc
library should probably be placed at or near the end of the library
list.

@item Enable the debugging options that you need by typing
@kbd{dmalloc -l logfile -i 100 low} (for example).  @kbd{dmalloc
--usage} will provide verbose usage info for the dmalloc program.
@xref{Dmalloc Program}.

@item Enable the ``lock-on'' option (for example @kbd{dmalloc -o 20}).
As explained above, you may have to try different values before getting
it right.  Values between 5 and 30 are probably good.

@item If you get a dmalloc error #13 @samp{thread locking has not
been configured} then you have not compiled you program with the
threaded version of dmalloc or there was a problem building it.

@item If everything works, you should be able to run your program, have
it not immediately crash, and the dmalloc library should not complain
about recursion.

@end enumerate

If you have any specific questions or would like addition information
posted in this section, please let me know.  Experienced thread
programmers only please.

@c --------------------------------

@node Using With Cygwin, Debugging A Server, Using With Threads, Programming
@section Using the library with Cygwin environment.

@cindex cygwin

The Cygwin environment is a Linux-like environment for Windows.  It
provides Linux look and feel as well as a programming environment.  See
URL @uref{http://www.cygwin.com/} for more details.

@cindex reading environment, problems with
@cindex getenv, problems with
@cindex GetEnvironmentVariableA, using with

Cygwin uses the @code{GetEnvironmentVariableA} function to read in
environmental variables instead of @code{getenv}.  This functions are
used to get the value of the @samp{DMALLOC_OPTIONS} variable which
sets the debugging options.  @xref{Environment Variable}.

@cindex HAVE_GETENVIRONMENTVARIABLEA
@cindex GETENV_SAVE conf.h option

As of right now, dmalloc is not detecting the
@code{GetEnvironmentVariableA} function correctly so you may need to
tune the @file{conf.h} file to get it to work.  See the sections on
@code{HAVE_GETENVIRONMENTVARIABLEA} and @code{GETENV_SAVE} settings.
Feedback is welcome here.

If you still have problems reading in the environmental variables, you
can work around this issue.  You can add some code into the
@code{main} function in your program to initialize the dmalloc flags
yourself.  Here is a code sample:

@cindex dmalloc_debug_setup usage

@example
main(int argc, char **argv)
@{
#ifdef DMALLOC
  /*
   * Get environ variable DMALLOC_OPTIONS and pass the settings string
   * on to dmalloc_debug_setup to setup the dmalloc debugging flags.
   */
  dmalloc_debug_setup(getenv("DMALLOC_OPTIONS"));
#endif

  /* rest of code in main starts here */
  @dots{}
@}
@end example

The @code{#ifdef} is just a good idea.  I means that when debugging with
dmalloc you need to compile your code with @code{-DDMALLOC}.  When you
are done debugging you can remove the flag and the call to
@code{dmalloc_debug_setup} will be removed.

Please let me know if there is a better way to do this.

@c --------------------------------

@node Debugging A Server, Logfile Details, Using With Cygwin, Programming
@section Debugging Memory in a Server or Cgi-Bin Process

@cindex cgi-bin process debugging
@cindex cgi-bin usage of dmalloc
@cindex child process debugging
@cindex daemon process debugging
@cindex debugging cgi-bin processes
@cindex debugging child processes
@cindex debugging daemon processes
@cindex debugging server processes
@cindex server process debugging
@cindex usage of dmalloc in a daemon
@cindex usage of dmalloc in a server
@cindex usage of dmalloc with cgi-bin

There are some specified challenges when trying to debug allocations
in processes which do not startup, run, and then shutdown.  Server
processes (often called daemons) are those that are started (often at
system boot time) and run perpetually.  Other processes which are
difficult to debug are CGI programs which are spawned by web servers
or when you want to start debugging inside of a child process.

@enumerate

@item Build your server or cgi-bin program with the dmalloc library
like any other program.  @xref{Getting Started}.

@item Add code into your program to enable the library flags to
perform the memory checks that you require.  Since these programs
often do not run from the command line, you cannot use the dmalloc
utility program and modify the process environment.  @xref{Dmalloc
Program}.  The library provides a couple of functions to set the
debugging flags when a program is running.

@cindex dmalloc_debug_setup function

@item To set the memory debugging flags, use the
@code{dmalloc_debug_setup} function which takes a string in the same
format of the @samp{DMALLOC_OPTIONS} environmental variable.
@xref{Environment Variable}.  Use the dmalloc utility with the
@code{-n} no-changes argument to see the appropriate settings for the
@samp{DMALLOC_OPTIONS} environmental variable.

@example
> dmalloc -n -l logfile high
Outputed:
DMALLOC_OPTIONS=debug=0x4f4ed03,log=logfile
export DMALLOC_OPTIONS
@end example

So if you want to turn on @kbd{high} debugging and log to the file
@file{logfile} then you would copy the above @samp{DMALLOC_OPTIONS}
value into a call to @code{dmalloc_debug_setup}.  Notice that I have
surrounded the dmalloc code with an @code{#ifdef DMALLOC} so you'll
have to compile using the @code{-DDMALLOC} flag.

@example
main()
@{
#ifdef DMALLOC
  /* set the 'high' flags */
  dmalloc_debug_setup("debug=0x4f47d03,log=logfile");
#endif
  @dots{}
@}
@end example

@emph{Please note} that the @code{dmalloc_debug_setup} function does
not know about @code{high}, @code{low}, or other debug tokens but
needs the actual flag values.

@item For earlier versions of the library (before 5.0.0) without 
@code{dmalloc_debug_setup}, the @code{dmalloc_debug} function is
available to set the flags directly, but it cannot adjust the logfile
name and the other environment settings.  You can use the dmalloc
utility program to see what the numerical equivalent of the @kbd{high}
token.

@example
> dmalloc -n high
Outputed:
DMALLOC_OPTIONS=debug=0x4f4ed03
export DMALLOC_OPTIONS
@end example

You can then take the @code{0x4f4ed03} hexadecimal number and call
@code{dmalloc_debug} with that number.

@example
main()
@{
#ifdef DMALLOC
  /* set the 'high' flags */
  dmalloc_debug(0x4f4ed03);
#endif
  @dots{}
@}
@end example

@item Even with the settings enabled, you may have problems getting
the logfile to be written if your program is running as @samp{nobody}
or another user without permissions for security reasons.  This is
especially true for cgi-bin programs.  In this case you should specify
a full path to your malloc logfile in a world writable directory
(ex. @code{dmalloc_debug_setup("debug=0x4f47d03,log=/var/tmp/malloc");}).
Watch for programs which change into other directories and which may
cause logfiles specified as relative or local paths to be dropped in
other locations.  You may always want to use a full path logfile.

@item Once you have your settings enabled and your log is being
generated, you may now want to check out how your process is doing in
terms of unfreed memory.  Since it is not shutting down, the automatic
unfreed log entries are not being dropped to the logfile.  By using
the @code{dmalloc_mark} and @code{dmalloc_log_changed} functions, you
can set a mark point at a certain place inside of your program, and
then later see whether there are any unfreed pointers since the mark.

@example
main()
@{
#ifdef DMALLOC
  /* set the 'high' flags */
  dmalloc_debug_setup("debug=0x4f47d03,log=logfile");
#endif

  while (1) @{
    /* accept a connection from a client */
    accept_connection();

    while (1) @{
#ifdef DMALLOC
      unsigned long mark;
      /* get the current dmalloc position */
      mark = dmalloc_mark() ;
#endif
      /* process the connection */
      if (process_connection() != PROCESS_OK) @{
        break;
      @}
#ifdef DMALLOC
      /*
       * log unfreed pointers that have been added to
       * the heap since mark
       */
      dmalloc_log_changed(mark,
                          1 /* log unfreed pointers */,
                          0 /* do not log freed pointers */,
                          1 /* log each pnt otherwise summary */);
#endif
    @}
    /* close the connection with the client */
    close_connection();
  @}
  @dots{}
@}
@end example

Usually you would set the mark after the initializations and before
each transaction is processed.  Then for each transaction you can use
@code{dmalloc_log_changed} to show the unfreed memory.
@xref{Extensions}.

@item You can also use the @code{dmalloc_log_stats} function to dump
general information about the heap.  Also, remember that you can use
the @code{dmalloc_message} and @code{dmalloc_vmessage} routines to
annotate the dmalloc logfile with details to help you debug memory
problems.  @xref{Extensions}.

@end enumerate

@c --------------------------------

@node Logfile Details, Other Hints, Debugging A Server, Programming
@section Explanation of the Logfile Output

@cindex logfile format

Most of time you will be using the logfile output from library as the
sole information source for diagnosing problems in and getting
statistics for your program.

@example
1098918225: 3: Dmalloc version '@value{dmalloc_version}'
1098918225: 3: flags = 0x4f4e503, logfile '/tmp/dmalloc.log'
1098918225: 3: interval = 500, addr = 0, seen # = 0, limit = 0
1098918225: 3: starting time = 1098918225
1098918225: 3: process pid = 32406
1098918226: 4: WARNING: tried to free(0) from foo.c:708'
1098918228: 20: *** free: at 'unknown' pnt '0xed310080|s2': \
            size 12, alloced at 'bar.c:102'
1098918230: 50: ERROR: heap_check: free space was overwritten (err 67)
1098918230: 50: error details: checking free pointer
1098918230: 50: pointer '0x291c5' from 'unknown' prev access 'foo.c:787'
@end example

Here is a short example of some logfile information.  Each of the
lines are prefixed by the time (in epoch seconds since 1/1/1970) and
the iteration or call count which is the number of times the library
has been called from malloc, free, verify, etc..  In the above
example, the first 5 log entries where written at epoch 1098918225 or
@samp{Wed Oct 27 19:03:45 2004 EST} and they were generated by the 3rd
call to the library.  See the @file{settings.dist} file entries to
tune what elements appear on each line: LOG_TIME_NUMBER,
LOG_ITERATION, LOG_PID, etc..  You can convert the epoch seconds to a
date from the command line with the following perl code: @code{perl -e
'print localtime($ARGV[0])."\n";' epoch-seconds-number}

@cindex version of library
The first 5 lines of the sample logfile contain header information for
all logfiles.  They show the version number and URL for the library as
well as all of the settings that the library is currently using.
These settings are tuned using the dmalloc utility program.
@xref{Dmalloc Program}.  The 5th line of is the process-id that
generated the logfile.

@cindex ALLOW_FREE_NULL_MESSAGE
The 6th line in the above example is what causes the logfile to be
opened and the header to be written.  It is a warning that tells you
that you tried to free a 0L pointer at a certain location.  You can
disable these warnings by setting @samp{ALLOW_FREE_NULL_MESSAGE} to 0
in @file{settings.dist}.

@cindex seen count
Line 7 is an example of a transaction log that you get when you enable
the @code{log-trans} debug token.  @xref{Debug Tokens}.  This line
shows that a call to free was made from an unknown location.  It is
unknown because the file in question did not include @file{dmalloc.h}
to get file/line-number information.  The call to free was freeing the
pointer address @code{0xed310080} which we have ``seen'' 2 times (s2).
We saw the pointer when it was allocated and then we are seeing it
again when it was freed.  Because the library is reusing pointers
(reclaiming freed memory) the seen count helps to track how many times
a pointer was used.  The last part of the line shows that the pointer
to be freed was allocated by @file{bar.c} line 102.

Lines 8-10 is the next problem that the library caught and this one is
an error.  It happened 5 seconds from the start of the log
(1098918230) and at the 50th call into the library.  It shows that an
allocation that had been freed then was overwritten.  This may imply
that someone tried to use memory after it was freed or that there was
a loose pointer reference.  The last two lines give more details about
when the error was discovered, the address of the offending pointer,
and when the pointer was previous accessed, in this case freed.  To
discover where this problem is happening, you can use a debugger.
@xref{Using With a Debugger}.

@c --------------------------------

@node Other Hints,, Logfile Details, Programming
@section Various Other Hints That May Help

@cindex other hints
@cindex hints that may help
@cindex various hints

One of the problems that is often seen is that a program crashes in
the @code{libc} memory code and you suspect a heap memory problem but
both dmalloc and maybe valgrind don't show any problems.  One of the
big problems with debugging is that it is very difficult to do it
without effecting how the program is run.  Sometimes errors are due to
subtle race conditions that are only seen when the program is running
at full speed -- not slowed down by debugging code.

This is especially true with threaded code which is often heavily
affected when used with dmalloc and valgrind.  Older versions of
valgrid (maybe current) forced all threads into a single virtual
system by design, which often masks reentrance bugs.

One way to work through these issues is to run with the library with
very few debugging flags enabled.  Many memory problems are fence-post
areas so start with dmalloc checking just the fence post and error
logging enabled:

@example
dmalloc -d 0 -l dmalloc.log -p log-stats -p log-non-free -p check-fence -p check-funcs 
@end example

This enabled a small number of checks and should cause your program to
run at close to full speed.  The library has never been optimized for
speed so some performance penalties will be felt.

@c ----------------------------------------------------------------------------

@node Dmalloc Program, Source Code, Programming, Top
@chapter Dmalloc Utility Program

@cindex dmalloc program
@cindex dmalloc utility
@cindex utility program
@cindex library utility

The dmalloc program is designed to assist in the setting of the
environment variable @samp{DMALLOC_OPTIONS}.  @xref{Environment
Variable}.  It is designed to print the shell commands necessary to make
the appropriate changes to the environment.  Unfortunately, it cannot
make the changes on its own so the output from dmalloc should be sent
through the @code{eval} shell command which will do the commands.

@menu
* Shell Alias::                 Using a shell alias with the utility.
* Utility Usage::               How to use the dmalloc program.
* Environment Variable::        Environment variable name and features.
* Debug Tokens::                Description of the debugging tokens.
* RC File::                     Format of the runtime configuration file.
@end menu

@c --------------------------------

@node Shell Alias, Utility Usage, Dmalloc Program, Dmalloc Program
@section Using a Shell Alias with the Utility

The dmalloc program is designed to assist in the setting of the
environment variable @samp{DMALLOC_OPTIONS}.  @xref{Environment
Variable}.  It is designed to print the shell commands necessary to make
the appropriate changes to the environment.  Unfortunately, it cannot
make the changes on its own so the output from dmalloc should be sent
through the @code{eval} shell command which will do the commands.

@cindex alias, shell
@cindex bash shell
@cindex ksh shell
@cindex zsh shell

With shells that have aliasing or macro capabilities: csh, bash, ksh,
tcsh, zsh, etc., setting up an alias to dmalloc to do the eval call is
recommended.  Bash, ksh, and zsh users should add the following to their
@file{.bashrc}, @file{.profile}, or @file{.zshrc} file respectively
(notice the @kbd{-b} option for bourne shell output):

@example
function dmalloc @{ eval `command dmalloc -b $*`; @}
@end example

If your shell does not support the @code{command} function then try:

@example
function dmalloc @{ eval `\dmalloc -b $*`; @}
@end example
or
@example
function dmalloc @{ eval `/usr/local/bin/dmalloc -b $*`; @}
@end example

@cindex tcsh shell
@cindex csh shell

If you are @emph{still} using csh or tcsh, you should add the following
to your @file{.cshrc} file (notice the @kbd{-C} option for c-shell
output):

@example
alias dmalloc 'eval `\dmalloc -C \!*`'
@end example

This allows the user to execute the dmalloc command as @samp{dmalloc
arguments}.

Users of versions of the Bourne shell (usually known as /bin/sh) that
don't have command functions will need to send the output to a temporary
file and the read it back in with the ``.'' command:

@example
$  dmalloc -b arguments @dots{} > /tmp/out
$  . /tmp/out
@end example

By the way, if you are looking for a shell, I heartily recommend trying
out zsh.  It is a bourne shell written from scratch with much the same
features as tcsh without the csh crap.

@emph{NOTE}: After you add the alias to the file you need to log out and
log back in to have it take effect, or you can execute the above
appropriate command on the command line.  If you enter @kbd{dmalloc
runtime} and see any output with DMALLOC_OPTIONS in it then the alias
did not work.

@c --------------------------------

@node Utility Usage, Environment Variable, Shell Alias, Dmalloc Program
@section How to Use the Dmalloc Program

@cindex utility usage
@cindex usage of the utility

The most basic usage for the program is @samp{dmalloc [-bC] tag}.  The
@samp{-b} or @samp{-C} (either but not both flags used at a time) are
for generating Bourne or C shell type commands respectively.  dmalloc
will try and use the @code{SHELL} environment variable to determine
whether bourne or C shell commands should be generated but you may want
to explicitly specify the correct flag.

The @samp{tag} argument to dmalloc should match a line from the user's
runtime configuration file or should be one of the built-in tags.
@xref{RC File}.  If no tag is specified and no other option-commands
used, dmalloc will display the current settings of the environment
variable.  It is useful to specify one of the verbose options when doing
this.

To find out the usage for the debug malloc program try @samp{dmalloc
--usage-long}.  The standardized usage message that will be displayed is
one of the many features of the argv library included with this package.

It is available on the web at URL @uref{http://256.com/sources/argv/}.
See the documentation there for more information.

Here is a detailed list of the flags that can passed to dmalloc:

@table @code
@item -a address
Set the @samp{addr} part of the @samp{DMALLOC_OPTIONS} variable to
address (or alternatively address:number).

@item -b
Output Bourne shell type commands.  Usually handled automagically.

@item -C
Output C shell type commands.  Usually handled automagically.

@item -c
Clear/unset all of the settings not specified with other arguments.  You
can do this automatically when you set to a new tag with the @kbd{-r}
option.

@emph{NOTE}: clear will never unset the @samp{debug} setting.  Use
@kbd{-d 0} or a tag to @samp{none} to achieve this.

@item -d bitmask
Set the @samp{debug} part of the @samp{DMALLOC_OPTIONS} env variable to
the bitmask value which should be in hex.  This is overridden (and
unnecessary) if a tag is specified.

@item -D
List all of the debug-tokens.  Useful for finding a token to be used
with the @kbd{-p} or @kbd{-m} options.  Use with @kbd{-v} or @kbd{-V}
verbose options.

@item -e errno
Print the dmalloc error string that corresponds to the error number
errno.

@item -f filename
Use this configuration file instead of the RC file
@file{$HOME/.dmallocrc}.

@item -g
Output gdb type commands for using inside of the gdb debugger.

@item -h (or --help)
Output a help message for the utility.

@item -i number
@cindex interval setting
Set the checking interval to number.  If the @code{check-heap} token is
enabled, this causes the library to only check the heap every Nth time
which can @emph{significantly} increase the running speed of your
program.  If a problem is found, however, this limits your ability to
determine when the problem occurred.  Try values of 50 or 100 initially.

@item -k
Do not reset all of the settings when a tag is specified.  This
specifically overrides the @kbd{-r} option and is provided here to
override @kbd{-r} if it has been added to the dmalloc alias.

@item -l filename
Write the debugging output and other log-file information to the
filename.  Filename can include some of the following patterns which get
expanded into strings:

@table @code

@cindex %h
@cindex gethostname function usage
@cindex hostname in logfile path
@cindex name of host in logfile path
@item %h
Gets expanded into the hostname if the @code{gethostname()} function is
available.

@cindex %i
@cindex thread-id in logfile path
@item %i
Gets expanded into the thread-id if the library has been configure to be
used with threads.  @xref{Using With Threads}.  See the end of the
@file{settings.dist} file for settings which return the thread-id and
convert it into a string.

@cindex %p
@cindex getpid function usage
@cindex pid in logfile path
@cindex process-id in logfile path
@item %p
Gets expanded into the process-id if the @code{getpid()} function is
available.

@cindex %t
@cindex time function usage
@cindex time in logfile path
@item %t
Gets expanded into the time value in seconds if the @code{time()}
function is available.

@cindex %u
@cindex getuid function usage
@cindex uid in logfile path
@cindex user-id in logfile path
@item %u
Gets expanded into the user-id number if the @code{getuid()} function is
available.

@end table

Some examples:

@example
# logfile produced with pid extension:
#   logfile.8412  or  logfile.31451
dmalloc -l logfile.%p

# hostname and time extensions:
#   dmalloc-box1.foo.com-1055213240
dmalloc -l dmalloc-%h-%t

# if threads enabled, have thread-id extension:  log.thread32
dmalloc -l log.thread%i
@end example

@item -L
Write the debug-value into the environment not in hex but by
individual debug-tokens in long form.

@item -m token(s)
Remove (minus) the debug capabilities of token(s) from the current debug
setting or from the selected tag (or @kbd{-d} value).  Multiple @kbd{-m}
options can be specified.

@cindex memory limit
@cindex ERROR_OVER_LIMIT
@item -M limit
Set the memory allocation limit which will abort the program if the
total memory allocations exceed this number of bytes.  The limit can
be a number with a k, m, or g at the end to indicate kilobyte,
megabyte, and gigabyte respectively.  Ex: 100k, 200m, 1g.  If the
limit is exceeded, this will generate an @code{ERROR_OVER_LIMIT}
error.  @xref{Error Codes}.

@item -n
Without changing the environment, output the commands resulting from the
supplied options.

@cindex lock on
@item -o times
Set the ``lock-on'' period which dictates to the threaded version of
the library to not initialize or lock the mutex lock around the library
until after a certain number of allocation calls have been made.  Some
number between 2 and 30 is probably good.  See the ``Using With
Threads'' section for more information about the operation of the
library with threads.  @xref{Using With Threads}.

@item -p token(s)
Add (plus) the debug capabilities of token(s) to the current debug
setting or to the selected tag (or @kbd{-d} value).  Multiple @kbd{-p}
options can be specified.

@item -r
Remove (unset) all settings when using a tag.  This is useful when you
are returning to a standard development tag and want the logfile,
address, and interval settings to be cleared automatically.  If you want
this behavior by default, this can be put into the dmalloc alias.

@item -R
Output rc shell type commands.  This is not for the runtime
configuration file but for the rc shell program.

@cindex delay heap checking
@cindex start heap check later
@item -s file:line
Set the @samp{start} part of the @samp{DMALLOC_OPTIONS} env variable to
a file-name and line-number location in the source where the library
should begin more extensive heap checking.  The file and line numbers
for heap transactions must be working for this option to be obeyed.
This is used if you are trying to locate a problem and you want the
extensive checking to not happen initially because it's too slow.

@cindex delay heap checking
@cindex start heap check later
@cindex LOG_ITERATION
@cindex interaction count
@cindex transaction count
@cindex mark count
@cindex memory transaction count
@item -S number
Set the @samp{start} part of the @samp{DMALLOC_OPTIONS} env variable to
an dmalloc mark number.  The library will begin more extensive heap
checking after this number of memory transactions.  If you
@code{LOG_ITERATION} enabled in your @file{settings.h} file then
the entries in the log file will be prepended with the number of memory
transactions that the library has handled so far.  This number can be
used to delay the start of the fine grained heap checking which can be
very slow.

@cindex delay heap checking
@cindex start heap check later
@item --start-size size
Set the @samp{start} part of the @samp{DMALLOC_OPTIONS} env variable to
a number of bytes.  The library will begin more extensive heap checking
after this amount of memory has been allocated by the library.  This
allows you to start the slow and detailed checking of the library later
in the program execution.  You can use patterns like 250m, 1g, or 102k
to mean 250 megabytes, 1 gigabyte, and 102 kilobytes respectively.

@item -t
List all of the tags in the rc-file.  Use with @kbd{-v} or @kbd{-V}
verbose options.

@item -u (or --usage)
Output the usage information for the utility.

@item -v
Give verbose output.  Especially useful when dumping current settings or
listing all of the tags.

@item -V
Give very verbose output for outputting even more details about
settings.

@cindex utility version
@cindex library version
@cindex version of utility
@item --version
Output the version string for the utility.  @emph{Please note} that
the version of the library that is installed or has been linked into
your application may be different from the utility version.

@end table

If no arguments are specified, dmalloc dumps out the current settings
that you have for the environment variable.  For example:

@example
Debug-Flags  '0x40005c7' (runtime)
Address      0x1f008, count = 3
Interval     100
Logpath      'malloc'
Start-File   not-set
@end example

With a -v option and no arguments, dmalloc dumps out the current
settings in a verbose manner.  For example:

@example
Debug-Flags  '0x40005c7' (runtime)
   log-stats, log-non-free, log-bad-space, check-fence, catch-null
Address      0x1f008, count = 10
Interval     100
Logpath      'malloc'
Start-File   not-set
@end example

Here are some examples of dmalloc usage:

@example
# start tough debugging, check the heap every 100 times,
# send the log information to file 'logfile'
dmalloc high -i 100 -l logfile

# find out what error code 20 is (from the logfile)
dmalloc -e 20

# cause the library to halt itself when it sees the address 0x34238
# for the 6th time.
dmalloc -a 0x34238:6

# send the log information to file 'logfile' with the time in seconds
# as an extension.
dmalloc -l logfile.%t

# return to the normal 'runtime' settings and clear out all
# other settings
dmalloc -c runtime

# enable basic 'low' settings plus (-p) the logging of
# transactions (log-trans) to file 'logfile'
dmalloc low -p log-trans -l logfile

# print out the current settings with Very-verbose output
dmalloc -V

# list the available debug malloc tokens with Very-verbose output
dmalloc -DV

# list the available tags from the rc file with verbose output
dmalloc -tv
@end example

@c --------------------------------

@node Environment Variable, Debug Tokens, Utility Usage, Dmalloc Program
@section Environment Variable Name and Features

@cindex environment variable
@cindex DMALLOC_OPTIONS

An @dfn{environment variable} is a variable that is part of the user's
working environment and is shared by all the programs.  The
@samp{DMALLOC_OPTIONS} variable is used by the dmalloc library to
enable or disable the memory debugging features, at runtime.
@emph{NOTE:} you can also use the @code{dmalloc_debug_setup} function
to set the option string.  It can be set either by hand or with the
help of the dmalloc program.  @xref{Dmalloc Program}.

@cindex shell usage

@cindex Bourne shell usage
@cindex sh usage
@cindex bash usage
@cindex ksh usage
@cindex zsh usage
To set it by hand, Bourne shell (sh, bash, ksh, or zsh) users should use:

@example
DMALLOC_OPTIONS=value
export DMALLOC_OPTIONS
@end example

@cindex C shell usage
@cindex csh usage
@cindex tcsh usage
C shell (csh or tcsh) users need to invoke:

@example
setenv DMALLOC_OPTIONS value
@end example

The value in the above examples is a comma separated list of tokens each
having a corresponding value.  The tokens are described below:

@table @code
@item debug
@cindex debug setting
This should be set to a value in hexadecimal which corresponds to the
functionality token values added together.  @xref{Debug Tokens}.  For
instance, if the user wanted to enable the logging of memory
transactions (value @samp{0x008}) and wanted to check fence-post memory
(value @samp{0x400}) then @samp{debug} should be set to @samp{0x408}
(@samp{0x008} + @samp{0x400}).

@emph{NOTE}: You don't have to worry about remembering all the hex
values of the tokens because the dmalloc program automates the setting
of this variable especially.

@cindex comma separated tokens in env variable

@emph{NOTE}: You can also specify the debug tokens directly, separated
by commas.  @xref{Debug Tokens}.  If @samp{debug} and the tokens are
both used, the token values will be added to the debug value.

@item lockon
@cindex lockon setting
Set this to a number which is the ``lock-on'' period.  This dictates
to the threaded version of the library to not initialize or lock the
mutex lock around the library until after a certain number of allocation
calls have been made.  See the ``Using With Threads'' section for more
information about the operation of the library with threads.
@xref{Using With Threads}.

@item log
@cindex logfile setting
@cindex logging information to disk
Set this to a filename so that if @samp{debug} has logging enabled, the
library can log transactions, administration information, and/or errors
to the file so memory problems and usage can be tracked.

To get different logfiles for different processes, you can assign
@samp{log} to a string with @code{%d} in it (for instance
@samp{logfile.%d}).  This will be replaced with the pid of the running
process (for instance @samp{logfile.2451}).

@emph{WARNING}: it is easy to core dump any program with dmalloc, if
you send in a format with arguments other than the one @code{%d}.

@item addr
@cindex address setting
@cindex address locating
@cindex tracking addresses
When this is set to a hex address (taken from the dmalloc log-file for
instance) dmalloc will abort when it finds itself either allocating or
freeing that address.

The address can also have an @samp{:number} argument.  For instance, if
it was set it to @samp{0x3e45:10}, the library will kill itself the 10th
time it sees address @samp{0x3e45}.  By setting the number argument to
0, the program will never stop when it sees the address.  This is useful
for logging all activity on the address and makes it easier to track
down specific addresses not being freed.

This works well in conjunction with the @code{STORE_SEEN_COUNT} option.
@xref{Memory Leaks}.

@emph{NOTE}: dmalloc will also log all activity on this address along
with a count.

@item inter
@cindex interval setting
By setting this to a number X, dmalloc will only check the heap every X
times.  This means a number of debugging features can be enabled while
still running the program within a finite amount of time.

A setting of @samp{100} works well with reasonably memory intensive
programs.  This of course means that the library will not catch errors
exactly when they happen but possibly 100 library calls later.

@item start
@cindex start setting
Set this to a number X and dmalloc will begin checking the heap after X
times.  This means the intensive debugging can be started after a
certain point in a program.

@samp{start} also has the format @samp{file:line}.  For instance, if it
is set to @samp{dmalloc_t.c:126} dmalloc will start checking the heap
after it sees a dmalloc call from the @file{dmalloc_t.c} file, line
number 126.  If you use @samp{dmalloc_t.c:0}, with a 0 line number, then
dmalloc will start checking the heap after it sees a call from anywhere
in the @file{dmalloc_t.c} file.

This allows the intensive debugging to be started after a certain
routine or file has been reached in the program.
@end table

Some examples are:

@example
# turn on transaction and stats logging and set
# 'logfile' as the log-file
setenv DMALLOC_OPTIONS log-trans,log-stats,log=logfile

# enable debug flags 0x1f as well as heap-checking and
# set the interval to be 100
setenv DMALLOC_OPTIONS debug=0x1f,check-heap,inter=100

# enable 'logfile' as the log-file, watch for
# address '0x1234', and start checking when we see
# file.c line 123
setenv DMALLOC_OPTIONS log=logfile,addr=0x1234,start=file.c:123
@end example

@c --------------------------------

@node Debug Tokens, RC File, Environment Variable, Dmalloc Program
@section Description of the Debugging Tokens

@cindex debug tokens
@cindex tokens, debug

The below tokens and their corresponding descriptions are for the
setting of the debug library setting in the environment variable.
@xref{Environment Variable}.  They should be specified in the user's
@file{.dmallocrc} file.  @xref{RC File}.

Each token, when specified, enables a specific debugging feature.  For
instance, if you have the @code{log-stats} token enabled, the library
will log general statistics to the logfile.

To get this information on the fly, use @kbd{dmalloc -DV}.  This will
print out the Debug tokens in Very-verbose mode.  @xref{Dmalloc
Program}.

@table @code
@cindex none token
@item none
No debugging functionality

@cindex log-stats
@item log-stats
Log general statistics when dmalloc_shutdown or dmalloc_log_stats is
called.

@cindex log-non-free
@item log-non-free
Log non-freed memory pointers when dmalloc_shutdown or dmalloc_log_unfreed
is called.

@cindex log-known
@item log-known
Log only known memory pointers that have not been freed.  Pointers which
do not have file/line or return-address information will not be logged.

@cindex log-trans
@item log-trans
Log general memory transactions (quite verbose).

@cindex log-admin
@item log-admin
Log administrative information (quite verbose).

@cindex log-bad-space
@item log-bad-space
Log actual bytes in and around bad pointers.

@cindex log-nonfree-space
@item log-nonfree-space
Log actual bytes in non-freed pointers.

@cindex log-elapsed-time
@item log-elapsed-time
Log elapsed-time for allocated pointers (see @file{conf.h}).

@cindex log-current-time
@item log-current-time
Log current-time for allocated pointers (see @file{conf.h}).

@cindex check-fence
@item check-fence
Check fence-post memory areas.

@cindex check-heap
@item check-heap
Verify heap administrative structure.

@cindex check-blank
@item check-blank
Check to see if space that was blanked when a pointer was allocated or
when it was freed has been overwritten.  If this is enabled then it
will enable @code{free-blank} and @code{alloc-blank} automatically.

@cindex check-funcs
@item check-funcs
Check the arguments of some functions (mostly string operations) looking
for bad pointers.

@cindex check-shutdown
@item check-shutdown
Check all of the pointers in the heap when the program exits.

@cindex catch-signals
@cindex signal shutdown
@cindex shutdown on signal
@cindex SIGHUP
@cindex SIGINT
@cindex SIGTERM
@cindex HUP signal
@cindex INT signal
@cindex TERM signal
@item catch-signals
Shutdown the library automatically on SIGHUP, SIGINT, or SIGTERM.  This
will cause the library to dump its statistics (if requested) when you
press control-c on the program (for example).

@cindex realloc-copy
@item realloc-copy
Always copy data to a new pointer when realloc.

@cindex blank space
@cindex blanking memory
@cindex overwriting memory
@cindex clearing memory
@cindex decimal 223 character
@cindex 223 character
@cindex octal 337 character
@cindex 337 character
@cindex 0337 character
@cindex hexadecimal 0xdf character
@cindex 0xdf character
@cindex df character
@cindex free-blank
@cindex ERROR_FREE_OVERWRITTEN
@item free-blank
Write special ``dmalloc-free'' byte (hexadecimal @code{0xdf}, octal
@code{0337}, decimal @code{223}) into space when it is freed.  You can
set this to be something else in the @file{settings.dist} file.  This
ensures that your program is not using memory after it has been freed.
You can check to see if areas have been improperly overwritten with
the @code{check-blank} token.  If the free space has been overwritten,
then @code{ERROR_FREE_OVERWRITTEN} is triggered.  @xref{Error Codes}.

@cindex dump core
@cindex core dump
@cindex error-abort
@item error-abort
Abort the program (and dump core) on errors.  See @code{error-dump}
below.  @xref{Dumping Core}.

@cindex blank space
@cindex blanking memory
@cindex overwriting memory
@cindex clearing memory
@cindex decimal 218 character
@cindex 218 character
@cindex octal 332 character
@cindex 332 character
@cindex 0332 character
@cindex hexadecimal 0xda character
@cindex 0xda character
@cindex da character
@cindex alloc-blank
@item alloc-blank
Write special ``dmalloc-alloc'' byte (hexadecimal @code{0xda}, octal
@code{0332}, decimal @code{218}) into space when it is allocated.  You
can set this to be something else in the @file{settings.dist} file.
If you are not using @code{calloc} this will overwrite the user space
with the special bytes ensuring that your program is initializing its
dynamic memory appropriately.  Also, if you ask for 35 bytes and the
library has to give you a block of 64 because of rounding issues, it
will overwrite the extra memory with the special byte.  You can then
check to see if the extra areas have been improperly overwritten by
enabling the @code{check-blank} token.

@cindex print-messages
@item print-messages
Log any errors and messages to the screen via standard-error.

@cindex mmap
@cindex sbrk
@cindex catch-null
@item catch-null
Abort the program immediately if the library fails to get more heap
space from the heap allocation routine @code{mmap} or @code{sbrk}.

@cindex never-reuse
@item never-reuse
Have the heap never use space that has been used before and freed.
@xref{Memory Leaks}.  @emph{WARNING}: This should be used with caution
since you may run out of heap space.

@cindex dump core
@cindex core dump
@cindex error-dump
@item error-dump
Dump core on error and then continue.  Later core dumps overwrite
earlier ones if the program encounters more than one error.  See
@code{error-abort} above.  @xref{Dumping Core}.

@emph{NOTE}: This will only work if your system supports the @code{fork}
system call and the configuration utility was able to fork without going
recursive.

@cindex error-free-null
@cindex ALLOW_FREE_NULL
@cindex ALLOW_FREE_NULL_MESSAGE

@item error-free-null
By default the library will not generate an error when a program tries
to free a NULL pointer.  By enabling this token, you can change this
behavior so an error is reported.  See also the ALLOW_FREE_NULL and
ALLOW_FREE_NULL_MESSAGE settings in the @file{settings.h} file to change
the default behavior.

@end table

@c --------------------------------

@node RC File,, Debug Tokens, Dmalloc Program
@section Format of the Runtime Configuration File

@cindex rc file
@cindex runtime-config file
@cindex configuration file
@cindex dmallocrc file
@cindex .dmallocrc file

By using a @dfn{RC File} (or runtime configuration file) you can alias
tags to combinations of debug tokens.  @xref{Debug Tokens}.

@emph{NOTE}: For beginning users, the dmalloc program has a couple of
tags built into it so it is not necessary for you to setup a RC file:

@table @code
@cindex runtime token
@cindex token runtime
@item runtime
Enables basic runtime tests including fence-post checking, null handling,
and logging of any errors.

@cindex low token
@cindex token low
@item low
Runtime settings plus minimal checking of heap structures and
overwriting of allocated and freed space.

@cindex medium token
@cindex token medium
@item medium
Low settings plus checking of all heap structures on each memory call,
always relocates block on realloc, and aborts on errors.  You may want
to use @kbd{-i} option to the dmalloc utility.  @xref{Dmalloc
Program}.

@cindex high token
@cindex token high
@item high
Medium settings plus checking of overwritten freed and allocated
memory and checking of arguments to a number of common functions. You
may want to use @kbd{-i} option to the dmalloc utility.  @xref{Dmalloc
Program}.

@end table

For expert users, a sample @file{dmallocrc} file has been provided but
you are encouraged to roll your own combinations.  The name of default
rc-file is @file{$HOME/.dmallocrc}.  The @samp{$HOME} environment
variable should be set by the system to point to your home-directory.

The file should contain lines in the general form of:

@example
tag     token1, token2, @dots{}
@end example

@samp{tag} is to be matched with the tag argument passed to the dmalloc
program, while @samp{token1, token2, @dots{}} are debug capability
tokens.  @xref{Dmalloc Program}, @ref{Debug Tokens}.

A line can be finished with a @samp{\} meaning it continues onto the
next line.  Lines beginning with @samp{#} are treated as comments and
are ignored along with empty lines.

Here is an example of a @file{.dmallocrc} file:

@example
#
# Dmalloc runtime configuration file for the debug malloc library
#

# no debugging
none    none

# basic debugging
debug1  log-stats, log-non-free, check-fence

# more logging and some heap checking
debug2  log-stats, log-non-free, log-trans, \
        check-fence, check-heap, error-abort

# good utilities
debug3  log-stats, log-non-free, log-trans, \
        log-admin, check-fence, check-heap, realloc-copy, \
        free-blank, error-abort

@dots{}
@end example

For example, with the above file installed, you can type @code{dmalloc
debug1} after setting up your shell alias.  @xref{Dmalloc Program}.
This enables the logging of statistics, the logging of non-freed memory,
and the checking of fence-post memory areas.

Enter @code{dmalloc none} to disable all memory debugging features.

@c ----------------------------------------------------------------------------

@node Source Code, Troubleshooting, Dmalloc Program, Top
@chapter Information on the Source Code

@cindex source code

@menu
* Definitions::                 Definition of terms and other information.
* Compatibility::               General compatibility concerns.
* Portability::                 Issues important for porting the library.
@end menu

@c --------------------------------

@node Definitions, Compatibility, Source Code, Source Code
@section Definition of Terms and other Information

@cindex source definitions

Here are a couple definitions and other information for those interested
in ``picking the brain'' of the library.  The code is a little ugly here
and there and it conforms to the Gray-Watson handbook of coding
standards only.

@table @dfn
@item bblock
basic block containing 2 ^ BASIC_BLOCK bytes of info

@item bblock_adm
administration for a set of basic blocks

@item dblock
divided block containing some base 2 number of blocks smaller than a
basic block.

@item dblock_adm
administration for a set of divided blocks

@item chunk
some anonymous amount of memory
@end table

For more information about administration structures, see the code and
comments from @file{chunk_loc.h}.

@c --------------------------------

@node Compatibility, Portability, Definitions, Source Code
@section General Compatibility Concerns

@cindex compatibility

@itemize @bullet
@item
Realloc() backwards compatibility with being able to realloc from the
last freed block is @emph{not} supported.  The author is interested to
know who is using this (cough, cough) feature and for what reason.

@cindex ALLOW_REALLOC_NULL settings.h option

@item
Realloc() of a NULL pointer is supported in which case the library
will just make a call to malloc().  This can be disabled with the help
of the @code{ALLOW_REALLOC_NULL} manual compilation option in the
@file{settings.h} file to adjust the library's default behavior.

@cindex ALLOW_FREE_NULL settings.h option

@item
Some systems allow free(0) to not be an error for some reason.  Since
0 is not a valid address returned by the malloc call, it is debatable
that this should be allowed.  See @file{settings.h} for the
@code{ALLOW_FREE_NULL} manual compilation option to adjust the
library's default behavior.

@item
Aside from possibly being slower than the system's memory allocation
functions, the library should be fully compatible with the standard
memory routines.  If this is @emph{not} the case, please bring this to
my attention.
@end itemize

@c --------------------------------

@node Portability,, Compatibility, Source Code
@section Issues Important for Porting the Library

@cindex portability

General portability issues center around:

@itemize @bullet
@item
@cindex mmap, usage without
@cindex sbrk, usage without
@cindex preallocated memory heap
@cindex INTERNAL_MEMORY_SPACE
mmap, sbrk, or compatible function usages.  The library does support a
preallocated memory chunk heap.  See the @code{INTERNAL_MEMORY_SPACE}
define in the @file{settings.dist} file.

@item
@cindex return-address
The locating of the caller's address from the dmalloc functions.
This is useful in locating problems from dmalloc functions called from C
files which did not include @file{dmalloc.h}: C library calls for
instance.

@cindex gcc
See @file{return.h} for the available architecture/compiler
combinations.  You may want to examine the assembly code from gcc (GNUs
superior c-compiler) version 2+ being run on the following code.  It
should give you a good start on building a hack for your box.

@example
static char * x;

a()
@{
        x = __builtin_return_address(0);
@}

main()
@{
        a();
@}
@end example

@end itemize

@c --------------------------------

@node Troubleshooting, Index of Concepts, Source Code, Top
@chapter Some Solutions to Common Problems

@cindex troubleshooting
@cindex common problems
@cindex help 
@cindex how do i...
@cindex faq
@cindex questions
@cindex problems

This section provides some answers to some common problems and
questions.  Please send me mail with any additions to
this list -- either problems you are still having or tips that you would
like to pass on.

When diagnosing a problem, if possible, always make sure you are running
the most up to date version of Dmalloc available from the home page at
URL @uref{http://dmalloc.com/}.  Problems are often fixed and a new
release can be published before people encounter them.

@table @samp

@cindex slow running
@cindex why running slow
@cindex too slow
@cindex why hanging
@cindex hanging program
@item Why does my program run so slow?

@cindex check-heap
This library has never been (and maybe never will be) optimized for
space nor speed.  Some of its features make it unable to use some of
the organizational methods of other more efficient heap libraries.

If you have the @code{check-heap} token enabled, your program might
run slow or seem to hang.  This is because by default, the library
will run a full check of the heap with every memory allocation or
free.  You can have the library check itself less frequently by using
the @kbd{-i} option to the dmalloc utility.  @xref{Dmalloc Program}.
If you are using the @kbd{high} token and you need your program to run
faster, try the @kbd{medium} or @kbd{low} tokens which don't check as
many heap features and so run faster although they will not catch as
many problems. @xref{RC File}.

@cindex logfile not produced
@cindex no logfile produced
@item Why was a log-file not produced after I ran my program?

This could be caused by a number of different problems.

@enumerate

@item Are you sure you followed all of the items in the ``Getting Started''
section?  Please review them if there is any doubt.  @xref{Getting
Started}.

@cindex env
@cindex printenv
@cindex DMALLOC_OPTIONS

@item Use the @kbd{env} or @kbd{printenv} commands to make sure that
the @samp{DMALLOC_OPTIONS} variable is set in your exported environment.
@xref{Environment Variable}.

@cindex ident
@cindex strings

@item Make sure that your program has been compiled correctly with the
dmalloc library.  The @kbd{ident} program should show chunk.c and other
dmalloc files compiled into your program.  You can also do @kbd{strings
-a your-program | grep chunk.c} and look for something like @samp{$Id:
chunk.c,v 1.152 1999/08/25 12:37:01 gray Exp $} with different versions
and date information.  If this doesn't show up then chances are dmalloc
was not linked into your program.

@item If your program changes its working directory, it may write the
dmalloc log-file somewhere else in the filesystem.  You will need to
check both where the program was started and to where it might change
directory.

@item The logfile is only produced when @code{dmalloc_shutdown()} is
called.  By default it will be called when @code{exit()} gets called.
If you are running your program and press @kbd{Control-C} under Unix
the program will stop immediately and @code{dmalloc_shutdown()} will
not get called.  You can either setup a signal handler for
@code{SIGINTR} and call exit yourself, or you can enable the
@code{catch-signals} token.  @xref{Debug Tokens}.

@item If your program is segfaulting or otherwise crashing when it exits, the
@code{exit()} routine may not being called.  You will have to resolve
these issues so the dmalloc library can gracefully exit and write its
log file.

@item You may want to call @code{dmalloc_log_stats()} and
@code{dmalloc_log_unfreed()} (or @code{dmalloc_log_changed()}) directly
to have the library write its log file.  Some system modules may not
have shutdown if you call this before @code{exit()} so extra unfreed
memory may be reported.

@end enumerate

@item I don't see any information about my non-freed (leaked) memory?

The library will not (by default) report on ``unknown'' non-freed
memory.  Unknown means memory that does not have associated file and
line information.

This will be necessary if you are @emph{not}
including @file{dmalloc.h} in all of your C files or if you are
interested in tracking leaks in system functions.

@item Dmalloc is returning the error "malloc library has gone recursive"

This most likely indicates that you are using the Dmalloc library within
a threaded application and two threads are trying to use the dmalloc
library at once.  Please see the section of the manual about threads for
more information about properly configuring the library.  @xref{Using
With Threads}.

If you are not using threads, then your program could have caught a
signal while within Dmalloc, which then in turn called a memory
allocation routine.  It is unwise to allocate memory on the heap in most
signal handlers.  Lastly, some functions called by the library may call
memory routines that it does not anticipate.  If you think this the
case, please report the problem and include a stack trace, operating
system version/type, and the version of Dmalloc you are using.

@end table

@c --------------------------------

@node Index of Concepts,, Troubleshooting, Top
@unnumbered Index of Concepts

@printindex cp
@contents

@bye