\input texinfo  @c -*-texinfo-*-
@setfilename gcc.info
@c @setfilename usegcc.info
@c To produce the full manual, use the "gcc.info" setfilename, and
@c make sure the following does NOT begin with '@c' (and the @clear line DOES)
@c To produce a user-only manual, use the "usegcc.info" setfilename, and
@c make sure the following does NOT begin with '@c':
@c @clear INTERNALS

@settitle Using and Porting GNU CC
@end ifset
@ifclear INTERNALS
@settitle Using GNU CC
@end ifclear

@syncodeindex fn cp
@syncodeindex vr cp

This file documents the use
and the internals
@end ifset
of the GNU compiler.

Copyright (C) 1988, 1989, 1992 Free Software Foundation, Inc.

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 process this file through Tex and print the
results, provided the printed document carries 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
sections entitled ``GNU General Public License'' and ``Protect Your
Freedom---Fight `Look And Feel'@w{}'' 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

Permission is granted to copy and distribute translations of this manual
into another language, under the above conditions for modified versions,
except that the sections entitled ``GNU General Public License'' and
``Protect Your Freedom---Fight `Look And Feel'@w{}'', and this permission
notice, may be included in translations approved by the Free Software
Foundation instead of in the original English.
@end ifinfo

@setchapternewpage odd

@center @titlefont{Using and Porting GNU CC}
@end ifset
@ifclear INTERNALS
@title Using GNU CC
@end ifclear
@sp 2
@center Richard M. Stallman
@sp 3
@center last updated 16 Dec 1992
@sp 1
@c The version number appears twice more in this file.
@center for version 2.3
@center (preliminary draft, which will change)
@vskip 0pt plus 1filll
Copyright @copyright{} 1988, 1989, 1992 Free Software Foundation, Inc.

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
sections entitled ``GNU General Public License'' and ``Boycott'' 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 sections entitled ``GNU General Public License'' and
``Boycott'', and this permission notice, may be included in translations
approved by the Free Software Foundation instead of in the original
@end titlepage


@node Top, Copying,, (DIR)
@top Introduction
@cindex introduction

This manual documents how to run, install and port the GNU C compiler, as
well as its new features and incompatibilities, and how to report bugs.
It corresponds to GNU CC version 2.3.
@end ifset

@ifclear INTERNALS
This manual documents how to run and install the GNU C compiler, as
well as its new features and incompatibilities, and how to report bugs.
It corresponds to GNU CC version 2.3.
@end ifclear

@end ifinfo
* Copying::         GNU General Public License says
                     how you can copy and share GNU CC.
* Contributors::    People who have contributed to GNU CC.
* Boycott::	    Protect your freedom---fight ``look and feel''.
* Invoking GCC::    Command options supported by @samp{gcc}.
* Installation::    How to configure, compile and install GNU CC.
* Extensions::      GNU extensions to the C language.
* Trouble::         If you have trouble installing GNU CC.
* Bugs::            How, why and where to report bugs.
* Service::         How to find suppliers of support for GNU CC.
* VMS::             Using GNU CC on VMS.
* Portability::     Goals of GNU CC's portability features.
* Interface::       Function-call interface of GNU CC output.
* Passes::          Order of passes, what they do, and what each file is for.
* RTL::             The intermediate representation that most passes work on.
* Machine Desc::    How to write machine description instruction patterns.
* Target Macros::   How to write the machine description C macros.
* Config::          Writing the @file{xm-@var{machine}.h} file.
* Index::	    Index of concepts and symbol names.
@end menu
@end ifset
@ifclear INTERNALS
* Copying::         GNU General Public License says
                     how you can copy and share GNU CC.
* Contributors::    People who have contributed to GNU CC.
* Boycott::	    Protect your freedom---fight ``look and feel''.
* Invoking GCC::    Command options supported by @samp{gcc}.
* Installation::    How to configure, compile and install GNU CC.
* Extensions::      GNU extensions to the C language.
* Trouble::         If you have trouble installing GNU CC.
* Bugs::            How, why and where to report bugs.
* Service::         How to find suppliers of support for GNU CC.
* VMS::             Using GNU CC on VMS.
* Index::	    Index of concepts and symbol names.
@end menu
@end ifclear

@node Copying, Contributors, Top, Top
@center Version 2, June 1991

Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
675 Mass Ave, Cambridge, MA 02139, USA

Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
@end display

@unnumberedsec Preamble

  The licenses for most software are designed to take away your
freedom to share and change it.  By contrast, the GNU General Public
License is intended to guarantee your freedom to share and change free
software---to make sure the software is free for all its users.  This
General Public License applies to most of the Free Software
Foundation's software and to any other program whose authors commit to
using it.  (Some other Free Software Foundation software is covered by
the GNU Library General Public License instead.)  You can apply it to
your programs, too.

  When we speak of free software, we are referring to freedom, not
price.  Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
this service if you wish), that you receive source code or can get it
if you want it, that you can change the software or use pieces of it
in new free programs; and that you know you can do these things.

  To protect your rights, we need to make restrictions that forbid
anyone to deny you these rights or to ask you to surrender the rights.
These restrictions translate to certain responsibilities for you if you
distribute copies of the software, or if you modify it.

  For example, if you distribute copies of such a program, whether
gratis or for a fee, you must give the recipients all the rights that
you have.  You must make sure that they, too, receive or can get the
source code.  And you must show them these terms so they know their

  We protect your rights with two steps: (1) copyright the software, and
(2) offer you this license which gives you legal permission to copy,
distribute and/or modify the software.

  Also, for each author's protection and ours, we want to make certain
that everyone understands that there is no warranty for this free
software.  If the software is modified by someone else and passed on, we
want its recipients to know that what they have is not the original, so
that any problems introduced by others will not reflect on the original
authors' reputations.

  Finally, any free program is threatened constantly by software
patents.  We wish to avoid the danger that redistributors of a free
program will individually obtain patent licenses, in effect making the
program proprietary.  To prevent this, we have made it clear that any
patent must be licensed for everyone's free use or not licensed at all.

  The precise terms and conditions for copying, distribution and
modification follow.

@end iftex
@end ifinfo

This License applies to any program or other work which contains
a notice placed by the copyright holder saying it may be distributed
under the terms of this General Public License.  The ``Program'', below,
refers to any such program or work, and a ``work based on the Program''
means either the Program or any derivative work under copyright law:
that is to say, a work containing the Program or a portion of it,
either verbatim or with modifications and/or translated into another
language.  (Hereinafter, translation is included without limitation in
the term ``modification''.)  Each licensee is addressed as ``you''.

Activities other than copying, distribution and modification are not
covered by this License; they are outside its scope.  The act of
running the Program is not restricted, and the output from the Program
is covered only if its contents constitute a work based on the
Program (independent of having been made by running the Program).
Whether that is true depends on what the Program does.

You may copy and distribute verbatim copies of the Program's
source code as you receive it, in any medium, provided that you
conspicuously and appropriately publish on each copy an appropriate
copyright notice and disclaimer of warranty; keep intact all the
notices that refer to this License and to the absence of any warranty;
and give any other recipients of the Program a copy of this License
along with the Program.

You may charge a fee for the physical act of transferring a copy, and
you may at your option offer warranty protection in exchange for a fee.

You may modify your copy or copies of the Program or any portion
of it, thus forming a work based on the Program, and copy and
distribute such modifications or work under the terms of Section 1
above, provided that you also meet all of these conditions:

@enumerate a
You must cause the modified files to carry prominent notices
stating that you changed the files and the date of any change.

You must cause any work that you distribute or publish, that in
whole or in part contains or is derived from the Program or any
part thereof, to be licensed as a whole at no charge to all third
parties under the terms of this License.

If the modified program normally reads commands interactively
when run, you must cause it, when started running for such
interactive use in the most ordinary way, to print or display an
announcement including an appropriate copyright notice and a
notice that there is no warranty (or else, saying that you provide
a warranty) and that users may redistribute the program under
these conditions, and telling the user how to view a copy of this
License.  (Exception: if the Program itself is interactive but
does not normally print such an announcement, your work based on
the Program is not required to print an announcement.)
@end enumerate

These requirements apply to the modified work as a whole.  If
identifiable sections of that work are not derived from the Program,
and can be reasonably considered independent and separate works in
themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works.  But when you
distribute the same sections as part of a whole which is a work based
on the Program, the distribution of the whole must be on the terms of
this License, whose permissions for other licensees extend to the
entire whole, and thus to each and every part regardless of who wrote it.

Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Program.

In addition, mere aggregation of another work not based on the Program
with the Program (or with a work based on the Program) on a volume of
a storage or distribution medium does not bring the other work under
the scope of this License.

You may copy and distribute the Program (or a work based on it,
under Section 2) in object code or executable form under the terms of
Sections 1 and 2 above provided that you also do one of the following:

@enumerate a
Accompany it with the complete corresponding machine-readable
source code, which must be distributed under the terms of Sections
1 and 2 above on a medium customarily used for software interchange; or,

Accompany it with a written offer, valid for at least three
years, to give any third party, for a charge no more than your
cost of physically performing source distribution, a complete
machine-readable copy of the corresponding source code, to be
distributed under the terms of Sections 1 and 2 above on a medium
customarily used for software interchange; or,

Accompany it with the information you received as to the offer
to distribute corresponding source code.  (This alternative is
allowed only for noncommercial distribution and only if you
received the program in object code or executable form with such
an offer, in accord with Subsection b above.)
@end enumerate

The source code for a work means the preferred form of the work for
making modifications to it.  For an executable work, complete source
code means all the source code for all modules it contains, plus any
associated interface definition files, plus the scripts used to
control compilation and installation of the executable.  However, as a
special exception, the source code distributed need not include
anything that is normally distributed (in either source or binary
form) with the major components (compiler, kernel, and so on) of the
operating system on which the executable runs, unless that component
itself accompanies the executable.

If distribution of executable or object code is made by offering
access to copy from a designated place, then offering equivalent
access to copy the source code from the same place counts as
distribution of the source code, even though third parties are not
compelled to copy the source along with the object code.

You may not copy, modify, sublicense, or distribute the Program
except as expressly provided under this License.  Any attempt
otherwise to copy, modify, sublicense or distribute the Program is
void, and will automatically terminate your rights under this License.
However, parties who have received copies, or rights, from you under
this License will not have their licenses terminated so long as such
parties remain in full compliance.

You are not required to accept this License, since you have not
signed it.  However, nothing else grants you permission to modify or
distribute the Program or its derivative works.  These actions are
prohibited by law if you do not accept this License.  Therefore, by
modifying or distributing the Program (or any work based on the
Program), you indicate your acceptance of this License to do so, and
all its terms and conditions for copying, distributing or modifying
the Program or works based on it.

Each time you redistribute the Program (or any work based on the
Program), the recipient automatically receives a license from the
original licensor to copy, distribute or modify the Program subject to
these terms and conditions.  You may not impose any further
restrictions on the recipients' exercise of the rights granted herein.
You are not responsible for enforcing compliance by third parties to
this License.

If, as a consequence of a court judgment or allegation of patent
infringement or for any other reason (not limited to patent issues),
conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License.  If you cannot
distribute so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you
may not distribute the Program at all.  For example, if a patent
license would not permit royalty-free redistribution of the Program by
all those who receive copies directly or indirectly through you, then
the only way you could satisfy both it and this License would be to
refrain entirely from distribution of the Program.

If any portion of this section is held invalid or unenforceable under
any particular circumstance, the balance of the section is intended to
apply and the section as a whole is intended to apply in other

It is not the purpose of this section to induce you to infringe any
patents or other property right claims or to contest validity of any
such claims; this section has the sole purpose of protecting the
integrity of the free software distribution system, which is
implemented by public license practices.  Many people have made
generous contributions to the wide range of software distributed
through that system in reliance on consistent application of that
system; it is up to the author/donor to decide if he or she is willing
to distribute software through any other system and a licensee cannot
impose that choice.

This section is intended to make thoroughly clear what is believed to
be a consequence of the rest of this License.

If the distribution and/or use of the Program is restricted in
certain countries either by patents or by copyrighted interfaces, the
original copyright holder who places the Program under this License
may add an explicit geographical distribution limitation excluding
those countries, so that distribution is permitted only in or among
countries not thus excluded.  In such case, this License incorporates
the limitation as if written in the body of this License.

The Free Software Foundation may publish revised and/or new versions
of the General Public License from time to time.  Such new versions will
be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.

Each version is given a distinguishing version number.  If the Program
specifies a version number of this License which applies to it and ``any
later version'', you have the option of following the terms and conditions
either of that version or of any later version published by the Free
Software Foundation.  If the Program does not specify a version number of
this License, you may choose any version ever published by the Free Software

If you wish to incorporate parts of the Program into other free
programs whose distribution conditions are different, write to the author
to ask for permission.  For software which is copyrighted by the Free
Software Foundation, write to the Free Software Foundation; we sometimes
make exceptions for this.  Our decision will be guided by the two goals
of preserving the free status of all derivatives of our free software and
of promoting the sharing and reuse of software generally.

@heading NO WARRANTY
@end iftex
@end ifinfo


@end enumerate

@end iftex
@end ifinfo

@unnumberedsec How to Apply These Terms to Your New Programs

  If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under these terms.

  To do so, attach the following notices to the program.  It is safest
to attach them to the start of each source file to most effectively
convey the exclusion of warranty; and each file should have at least
the ``copyright'' line and a pointer to where the full notice is found.

@var{one line to give the program's name and an idea of what it does.}
Copyright (C) 19@var{yy}  @var{name of author}

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
@end smallexample

Also add information on how to contact you by electronic and paper mail.

If the program is interactive, make it output a short notice like this
when it starts in an interactive mode:

Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
type `show w'.  This is free software, and you are welcome
to redistribute it under certain conditions; type `show c'
for details.
@end smallexample

The hypothetical commands @samp{show w} and @samp{show c} should show
the appropriate parts of the General Public License.  Of course, the
commands you use may be called something other than @samp{show w} and
@samp{show c}; they could even be mouse-clicks or menu items---whatever
suits your program.

You should also get your employer (if you work as a programmer) or your
school, if any, to sign a ``copyright disclaimer'' for the program, if
necessary.  Here is a sample; alter the names:

Yoyodyne, Inc., hereby disclaims all copyright
interest in the program `Gnomovision'
(which makes passes at compilers) written 
by James Hacker.

@var{signature of Ty Coon}, 1 April 1989
Ty Coon, President of Vice
@end group
@end example

This General Public License does not permit incorporating your program into
proprietary programs.  If your program is a subroutine library, you may
consider it more useful to permit linking proprietary applications with the
library.  If this is what you want to do, use the GNU Library General
Public License instead of this License.

@node Contributors, Boycott, Copying, Top
@unnumbered Contributors to GNU CC
@cindex contributors

In addition to Richard Stallman, several people have written parts
of GNU CC.

@itemize @bullet
The idea of using RTL and some of the optimization ideas came from the
program PO written at the University of Arizona by Jack Davidson and
Christopher Fraser.  See ``Register Allocation and Exhaustive Peephole
Optimization'', Software Practice and Experience 14 (9), Sept. 1984,

Paul Rubin wrote most of the preprocessor.

Leonard Tower wrote parts of the parser, RTL generator, and RTL
definitions, and of the Vax machine description.

Ted Lemon wrote parts of the RTL reader and printer.

Jim Wilson implemented loop strength reduction and some other
loop optimizations.

Nobuyuki Hikichi of Software Research Associates, Tokyo, contributed
the support for the Sony NEWS machine.

Charles LaBrec contributed the support for the Integrated Solutions
68020 system.

Michael Tiemann of Cygnus Support wrote the front end for C++, as well
as the support for inline functions and instruction scheduling.  Also
the descriptions of the National Semiconductor 32000 series cpu, the
SPARC cpu and part of the Motorola 88000 cpu.

Jan Stein of the Chalmers Computer Society provided support for
Genix, as well as part of the 32000 machine description.

Randy Smith finished the Sun FPA support.

Robert Brown implemented the support for Encore 32000 systems.

David Kashtan of SRI adapted GNU CC to the Vomit-Making System (VMS).

Alex Crain provided changes for the 3b1.

Greg Satz and Chris Hanson assisted in making GNU CC work on HP-UX for
the 9000 series 300.

William Schelter did most of the work on the Intel 80386 support.

Christopher Smith did the port for Convex machines.

Paul Petersen wrote the machine description for the Alliant FX/8.

Alain Lichnewsky ported GNU CC to the Mips cpu.

Devon Bowen, Dale Wiles and Kevin Zachmann ported GNU CC to the Tahoe.

Jonathan Stone wrote the machine description for the Pyramid computer.

Gary Miller ported GNU CC to Charles River Data Systems machines.

Richard Kenner of the New York University Ultracomputer Research
Laboratory wrote the machine descriptions for the AMD 29000, the DEC
Alpha, the IBM RT PC, and the IBM RS/6000 as well as the support for
instruction attributes.  He also made changes to better support RISC
processors including changes to common subexpression elimination,
strength reduction, function calling sequence handling, and condition
code support, in addition to generalizing the code for frame pointer

Richard Kenner and Michael Tiemann jointly developed reorg.c, the delay
slot scheduler.

Mike Meissner and Tom Wood of Data General finished the port to the
Motorola 88000.

Masanobu Yuhara of Fujitsu Laboratories implemented the machine
description for the Tron architecture (specifically, the Gmicro).

NeXT, Inc.@: donated the front end that supports the Objective C
@c We need to be careful to make it clear that "Objective C"
@c is the name of a language, not that of a program or product.

James van Artsdalen wrote the code that makes efficient use of
the Intel 80387 register stack.

Mike Meissner at the Open Software Foundation finished the port to the
MIPS cpu, including adding ECOFF debug support.

Ron Guilmette implemented the @code{protoize} and @code{unprotoize}
tools, the support for Dwarf symbolic debugging information, and much of
the support for System V Release 4.  He has also worked heavily on the
Intel 386 and 860 support.

Torbjorn Granlund of the Swedish Institute of Computer Science
implemented multiply-by-constant optimization and better long long
support, and improved leaf function register allocation.

Mike Stump implemented the support for Elxsi 64 bit CPU.
@end itemize

@node Boycott
@chapter Protect Your Freedom---Fight ``Look And Feel''

@i{This section is a political message from the League for Programming
Freedom to the users of GNU CC.  It is included here as an expression
of support for the League on the part of the Free Software Foundation.}
@end quotation

Apple and Lotus are trying to create a new form of legal monopoly: a
copyright on a class of user interfaces.  These monopolies would cause
serious problems for users and developers of computer software and
systems.  Xerox, too, has tried to make a monopoly for itself on window
systems; their suit against Apple was thrown out on a technicality, but
Xerox has not said anything to indicate it wouldn't try again.

Until a few years ago, the law seemed clear: no one could restrict
others from using a user interface; programmers were free to implement
any interface they chose.  Imitating interfaces, sometimes with changes,
was standard practice in the computer field.  The interfaces we know
evolved gradually in this way; for example, the Macintosh user interface
drew ideas from the Xerox interface, which in turn drew on work done at
Stanford and SRI.  1-2-3 imitated VisiCalc, and dBase imitated a
database program from JPL.

Most computer companies, and nearly all computer users, were happy with
this state of affairs.  The companies that are suing say it does not
offer ``enough incentive'' to develop their products, but they must have
considered it ``enough'' when they made their decision to do so.  It
seems they are not satisfied with the opportunity to continue to compete
in the marketplace---not even with a head start.

If companies like Xerox, Lotus, and Apple are permitted to make law
through the courts, the precedent will hobble the software industry:

@itemize @bullet
Gratuitous incompatibilities will burden users.  Imagine if each
car manufacturer had to arrange the pedals in a different order.

Software will become and remain more expensive.  Users will be
``locked in'' to proprietary interfaces, for which there is no real

Large companies have an unfair advantage wherever lawsuits become
commonplace.  Since they can easily afford to sue, they can intimidate
small companies with threats even when they don't really have a case.

User interface improvements will come slower, since incremental
evolution through creative imitation will no longer be permitted.

Even Apple, etc., will find it harder to make improvements if
they can no longer adapt the good ideas that others introduce, for
fear of weakening their own legal positions.  Some users suggest that
this stagnation may already have started.

If you use GNU software, you might find it of some concern that user
interface copyright will make it hard for the Free Software Foundation
to develop programs compatible with the interfaces that you already
@end itemize

To protect our freedom from lawsuits like these, a group of programmers
and users have formed a new grass-roots political organization, the
League for Programming Freedom.

The purpose of the League is to oppose new monopolistic practices such
as user-interface copyright and software patents; it calls for a return
to the legal policies of the recent past, in which these practices were
not allowed.  The League is not concerned with free software as an
issue, and not affiliated with the Free Software Foundation.

The League's membership rolls include John McCarthy, inventor of Lisp,
Marvin Minsky, founder of the Artificial Intelligence lab, Guy L.
Steele, Jr., author of well-known books on Lisp and C, as well as
Richard Stallman, the developer of GNU CC.  Please join and add your
name to the list.  Membership dues in the League are $42 per year for
programmers, managers and professionals; $10.50 for students; $21 for

The League needs both activist members and members who only pay their

To join, or for more information, phone (617) 243-4091 or write to:

League for Programming Freedom
1 Kendall Square #143
P.O. Box 9171
Cambridge, MA 02139
@end display

You can also send electronic mail to @code{league@@prep.ai.mit.edu}.

Here are some suggestions from the League for things you can do to
protect your freedom to write programs:

@itemize @bullet
Don't buy from Xerox, Lotus or Apple.  Buy from their competitors or
from the defendants they are suing.

Don't develop software to work with the systems made by these companies.

Port your existing software to competing systems, so that you encourage
users to switch.

Write letters to company presidents to let them know their conduct
is unacceptable.

Tell your friends and colleagues about this issue and how it threatens
to ruin the computer industry.

Above all, don't work for the look-and-feel plaintiffs, and don't
accept contracts from them.

Write to Congress to explain the importance of this issue.

House Subcommittee on Intellectual Property
2137 Rayburn Bldg
Washington, DC 20515

Senate Subcommittee on Patents, Trademarks and Copyrights
United States Senate
Washington, DC 20510
@end display

(These committees have received lots of mail already; let's give them
even more.)
@end itemize

Express your opinion!  You can make a difference.

@include invoke.texi

@include install.texi

@include extend.texi

@node Trouble
@chapter Known Causes of Trouble with GNU CC
@cindex bugs, known
@cindex installation trouble
@cindex known causes of trouble

This section describes known problems that affect users of GNU CC.  Most
of these are not GNU CC bugs per se---if they were, we would fix them.
But the result for a user may be like the result of a bug.

Some of these problems are due to bugs in other software, some are
missing features that are too much work to add, and some are places
where people's opinions differ as to what is best.

* Actual Bugs::		      Bugs we will fix later.
* Installation Problems::     Problems that manifest when you install GNU CC.
* Cross-Compiler Problems::   Common problems of cross compiling with GNU CC.
* Interoperation::      Problems using GNU CC with other compilers,
			   and with certain linkers, assemblers and debuggers.
* Incompatibilities::   GNU CC is incompatible with traditional C.
* Disappointments::     Regrettable things we can't change, but not quite bugs.
* Protoize Caveats::    Things to watch out for when using @code{protoize}.
* Non-bugs::		Things we think are right, but some others disagree.
@end menu

@node Actual Bugs
@section Actual Bugs We Haven't Fixed Yet

@itemize @bullet
Loop unrolling doesn't work properly for certain C++ programs.  This is
because of difficulty in updating the debugging information within the
loop being unrolled.  We plan to revamp the representation of debugging
information so that this will work properly, but we have not done this
in version 2.3 because we don't want to delay it any further.
@end itemize

@node Installation Problems
@section Installation Problems

This is a list of problems (and some apparent problems which don't
really mean anything is wrong) that show up during installation of GNU

@itemize @bullet
On certain systems, defining certain environment variables such as
@code{CC} can interfere with the functioning of @code{make}.

If you encounter seemingly strange errors when trying to build the
compiler in a directory other than the source directory, it could be
because you have previously configured the compiler in the source
directory.  Make sure you have done all the necessary preparations.
@xref{Other Dir}.

In previous versions of GNU CC, the @code{gcc} driver program looked for
@code{as} and @code{ld} in various places such as files beginning with
@file{/usr/local/lib/gcc-}.  GNU CC version 2 looks for them in the
directory @file{/usr/local/lib/gcc-lib/@var{target}/@var{version}}.

Thus, to use a version of @code{as} or @code{ld} that is not the system
default, for example @code{gas} or GNU @code{ld}, you must put them in
that directory (or make links to them from that directory).

Some commands executed when making the compiler may fail (return a
non-zero status) and be ignored by @code{make}.  These failures, which
are often due to files that were not found, are expected, and can safely
be ignored.

It is normal to have warnings in compiling certain files about
unreachable code and about enumeration type clashes.  These files' names
begin with @samp{insn-}.

Sometimes @code{make} recompiles parts of the compiler when installing
the compiler.  In one case, this was traced down to a bug in
@code{make}.  Either ignore the problem or switch to GNU Make.

On some 386 systems, building the compiler never finishes because
@code{enquire} hangs due to a hardware problem in the motherboard---it
reports floating point exceptions to the kernel incorrectly.  You can
install GNU CC except for @file{float.h} by patching out the command to
run @code{enquire}.  You may also be able to fix the problem for real by
getting a replacement motherboard.  This problem was observed in
Revision E of the Micronics motherboard, and is fixed in Revision F.

On some 386 systems, GNU CC crashes trying to compile @file{enquire.c}.
This happens on machines that don't have a 387 FPU chip.  On 386
machines, the system kernel is supposed to emulate the 387 when you
don't have one.  The crash is due to a bug in the emulator.

One of these systems is the Unix from Interactive Systems: 386/ix.
On this system, an alternate emulator is provided, and it does work.
To use it, execute this command as super-user:

ln /etc/emulator.rel1 /etc/emulator
@end example

and then reboot the system.  (The default emulator file remains present
under the name @file{emulator.dflt}.)

If you have such a problem on the SCO system, try using

Another system which has this problem is Esix.  We don't know whether it
has an alternate emulator that works.

@cindex @code{genflags}, crash on Sun 4
Sometimes on a Sun 4 you may observe a crash in the program
@code{genflags} or @code{genoutput} while building GNU CC.  This is said to
be due to a bug in @code{sh}.  You can probably get around it by running
@code{genflags} or @code{genoutput} manually and then retrying the

If you use the 1.31 version of the MIPS assembler (such as was shipped
with Ultrix 3.1), you will need to use the -fno-delayed-branch switch
when optimizing floating point code.  Otherwise, the assembler will
complain when the GCC compiler fills a branch delay slot with a
floating point instruction, such as add.d.

Users have reported some problems with version 2.0 of the MIPS
compiler tools that were shipped with Ultrix 4.1.  Version 2.10
which came with Ultrix 4.2 seems to work fine.

Some versions of the MIPS linker will issue an assertion failure
when linking code that uses @code{alloca} against shared
libraries on RISC-OS 5.0, and DEC's OSF/1 systems.  This is a bug
in the linker, that is supposed to be fixed in future revisions.
To protect against this, GCC passes @samp{-non_shared} to the
linker unless you pass an explicit @samp{-shared} or
@samp{-call_shared} switch.

On System V release 3, you may get this error message
while linking:

ld fatal: failed to write symbol name @var{something} 
 in strings table for file @var{whatever}
@end smallexample

This indicates that the disk is full or your ULIMIT won't allow
the file to be as large as it needs to be.

On HP 9000 series 300 or 400 running HP-UX release 8.0, there is a bug
in the assembler that must be fixed before GNU CC can be built.  This
bug manifests itself during the first stage of compilation, while
building @file{libgcc2.a}:

cc1: warning: `-g' option not supported on this version of GCC
cc1: warning: `-g1' option not supported on this version of GCC
./gcc: Internal compiler error: program as got fatal signal 11
@end example

A patched version of the assembler is available by anonymous ftp from
@code{altdorf.ai.mit.edu} as the file
@file{archive/cph/hpux-8.0-assembler}.  If you have HP software support,
the patch can also be obtained directly from HP, as described in the
following note:

This is the patched assembler, to patch SR#1653-010439, where the
assembler aborts on floating point constants.

The bug is not really in the assembler, but in the shared library
version of the function ``cvtnum(3c)''.  The bug on ``cvtnum(3c)'' is
SR#4701-078451.  Anyway, the attached assembler uses the archive
library version of ``cvtnum(3c)'' and thus does not exhibit the bug.
@end quotation

This patch is also known as PHCO_0800.

Another assembler problem on the HP PA results in an error message like
this while compiling part of @file{libgcc2.a}:

as: /usr/tmp/cca08196.s @@line#30 [err#1060]
  Argument 1 or 3 in FARG upper
         - lookahead = RTNVAL=GR
@end example

This happens because HP changed the assembler syntax after system
release 8.02.  GNU CC assumes the newer syntax; if your assembler wants
the older syntax, comment out this line in the file @file{pa1-hpux.h}:

@end example

Some versions of the Pyramid C compiler are reported to be unable to
compile GNU CC.  You must use an older version of GNU CC for
bootstrapping.  One indication of this problem is if you get a crash
when GNU CC compiles the function @code{muldi3} in file @file{libgcc2.c}.

You may be able to succeed by getting GNU CC version 1, installing it,
and using it to compile GNU CC version 2.  The bug in the Pyramid C
compiler does not seem to affect GNU CC version 1.

On the Tower models 4@var{n}0 and 6@var{n}0, by default a process is not
allowed to have more than one megabyte of memory.  GNU CC cannot compile
itself (or many other programs) with @samp{-O} in that much memory.

To solve this problem, reconfigure the kernel adding the following line
to the configuration file:

MAXUMEM = 4096
@end example

On the Altos 3068, programs compiled with GNU CC won't work unless you
fix a kernel bug.  This happens using system versions V.2.2 1.0gT1 and
V.2.2 1.0e and perhaps later versions as well.  See the file

You will get several sorts of compilation and linking errors on the
we32k if you don't follow the special instructions.  @xref{WE32K
@end itemize

@node Cross-Compiler Problems
@section Cross-Compiler Problems

@itemize @bullet
Cross compilation can run into trouble for certain machines because
some target machines' assemblers require floating point numbers to be
written as @emph{integer} constants in certain contexts.

The compiler writes these integer constants by examining the floating
point value as an integer and printing that integer, because this is
simple to write and independent of the details of the floating point
representation.  But this does not work if the compiler is running on
a different machine with an incompatible floating point format, or
even a different byte-ordering.

In addition, correct constant folding of floating point values
requires representing them in the target machine's format.
(The C standard does not quite require this, but in practice
it is the only way to win.)

It is now possible to overcome these problems by defining macros such
as @code{REAL_VALUE_TYPE}.  But doing so is a substantial amount of
work for each target machine.  @xref{Cross-compilation}.
@end ifset
@ifclear INTERNALS
It is now possible to overcome these problems by defining macros such
as @code{REAL_VALUE_TYPE}.  But doing so is a substantial amount of
work for each target machine.  @xref{Cross-compilation,,Cross
Compilation and Floating Point Format, gcc.info, Using and Porting GCC}.
@end ifclear

At present, the program @file{mips-tfile} which adds debug
support to object files on MIPS systems does not work in a cross
compile environment.
@end itemize

@node Interoperation
@section Interoperation

This section lists various difficulties encountered in using GNU C or
GNU C++ together with other compilers or with the assemblers, linkers,
libraries and debuggers on certain systems.

@itemize @bullet
GNU C normally compiles functions to return small structures and unions
in registers.  Most other compilers arrange to return them just like
larger structures and unions.  This can lead to trouble when you link
together code compiled by different compilers. To avoid the problem, you
can use the option @samp{-fpcc-struct-return} when compiling with GNU CC.

GNU C++ does not do name mangling in the same way as other C++
compilers.  This means that object files compiled with one compiler
cannot be used with another.

GNU C++ also uses different techniques for arranging virtual function
tables and the layout of class instances.  In general, therefore,
linking code compiled with different C++ compilers does not work.

Older GDB versions sometimes fail to read the output of GNU CC version
2.  If you have trouble, get GDB version 4.4 or later.

@cindex DBX
DBX rejects some files produced by GNU CC, though it accepts similar
constructs in output from PCC.  Until someone can supply a coherent
description of what is valid DBX input and what is not, there is
nothing I can do about these problems.  You are on your own.

The GNU assembler (GAS) does not support PIC.  To generate PIC code, you
must use some other assembler, such as @file{/bin/as}.

On some BSD systems including some versions of Ultrix, use of profiling
causes static variable destructors (currently used only in C++) not to
be run.

Use of @samp{-I/usr/include} may cause trouble.

Many systems come with header files that won't work with GNU CC unless
corrected by @code{fixincludes}.  The corrected header files go in a new
directory; GNU CC searches this directory before @file{/usr/include}.
If you use @samp{-I/usr/include}, this tells GNU CC to search
@file{/usr/include} earlier on, before the corrected headers.  The
result is that you get the uncorrected header files.

Instead, you should use these options:

-I/usr/local/lib/gcc-lib/@var{target}/@var{version}/include -I/usr/include
@end example

@cindex @code{vfork}, for the Sun-4
There is a bug in @code{vfork} on the Sun-4 which causes the registers
of the child process to clobber those of the parent.  Because of this,
programs that call @code{vfork} are likely to lose when compiled
optimized with GNU CC when the child code alters registers which contain
C variables in the parent.  This affects variables which are live in the
parent across the call to @code{vfork}.

If you encounter this, you can work around the problem by declaring
variables @code{volatile} in the function that calls @code{vfork}, until
the problem goes away, or by not declaring them @code{register} and not
using @samp{-O} for those source files.
@end ignore

On a Sparc, GNU CC aligns all values of type @code{double} on an 8-byte
boundary, and it expects every @code{double} to be so aligned.  The Sun
compiler usually gives @code{double} values 8-byte alignment, with one
exception: function arguments of type @code{double} may not be aligned.

As a result, if a function compiled with Sun CC takes the address of an
argument of type @code{double} and passes this pointer of type
@code{double *} to a function compiled with GNU CC, dereferencing the
pointer may cause a fatal signal.

One way to solve this problem is to compile your entire program with GNU
CC.  Another solution is to modify the function that is compiled with
Sun CC to copy the argument into a local variable; local variables
are always properly aligned.  A third solution is to modify the function
that uses the pointer to dereference it via the following function
@code{access_double} instead of directly with @samp{*}:

inline double
access_double (double *unaligned_ptr)
  union d2i @{ double d; int i[2]; @};

  union d2i *p = (union d2i *) unaligned_ptr;
  union d2i u;

  u.i[0] = p->i[0];
  u.i[1] = p->i[1];

  return u.d;
@end example

Storing into the pointer can be done likewise with the same union.

On a Sun, linking using GNU CC fails to find a shared library and
reports that the library doesn't exist at all.

This happens if you are using the GNU linker, because it does only
static linking and looks only for unshared libraries.  If you have a
shared library with no unshared counterpart, the GNU linker won't find

We hope to make a linker which supports Sun shared libraries, but please
don't ask when it will be finished---we don't know.

Sun forgot to include a static version of @file{libdl.a} with some
versions of SunOS (mainly 4.1).  This results in undefined symbols when
linking static binaries (that is, if you use @samp{-static}).  If you
see undefined symbols @code{_dlclose}, @code{_dlsym} or @code{_dlopen}
when linking, compile and link against the file
@file{mit/util/misc/dlsym.c} from the MIT version of X windows.

On the HP PA machine, ADB sometimes fails to work on functions compiled
with GNU CC.  Specifically, it fails to work on functions that use
@code{alloca} or variable-size arrays.  This is because GNU CC doesn't
generate HP-UX unwind descriptors for such functions.  It may even be
impossible to generate them.

Debugging (@samp{-g}) is not supported on the HP PA machine, unless you use 
the preliminary GNU tools (@pxref{Installation}).

The HP-UX linker has a bug which can cause programs which make use of
@code{const} variables to fail in unusual ways.  If your program makes
use of global @code{const} variables, we suggest you compile with the
following additional options:

-Dconst="" -D__const="" -D__const__="" -fwritable-strings
@end example

This will force the @code{const} variables into the DATA subspace which 
will avoid the linker bug.

Another option one might use to work around this problem is
@samp{-mkernel}.  @samp{-mkernel} changes how the address of variables
is computed to a sequence less likely to tickle the HP-UX linker bug.

We hope to work around this problem in GNU CC 2.4, if HP does not fix

Taking the address of a label may generate errors from the HP-UX
PA assembler.  GAS for the PA does not have this problem.

GNU CC produced code will not yet link against HP-UX 8.0 shared libraries.
We expect to fix this problem in GNU CC 2.4.  

The current version of the assembler (@file{/bin/as}) for the RS/6000
has certain problems that prevent the @samp{-g} option in GCC from

IBM has produced a fixed version of the assembler.  The replacement
assembler is not a standard component of either AIX 3.1.5 or AIX 3.2,
but is expected to become standard in a future distribution.  This
assembler is available from IBM as APAR IX22829.  Yet more bugs have
been fixed in a newer assembler, which will shortly be available as APAR
IX26107.  See the file @file{README.RS6000} for more details on these

On the IBM RS/6000, compiling code of the form

extern int foo;

@dots{} foo @dots{}

static int foo;
@end example

will cause the linker to report an undefined symbol @code{foo}.
Although this behavior differs from most other systems, it is not a
bug because redefining an @code{extern} variable as @code{static}
is undefined in ANSI C.

On VMS, GAS versions 1.38.1 and earlier may cause spurious warning
messages from the linker.  These warning messages complain of mismatched
psect attributes.  You can ignore them.  @xref{VMS Install}.

On NewsOS version 3, if you include both @file{stddef.h} and
@file{sys/types.h}, you get an error because there are two typedefs of
@code{size_t}.  You should change @file{sys/types.h} by adding these
lines around the definition of @code{size_t}:

#ifndef _SIZE_T
#define _SIZE_T
@var{actual typedef here}
@end example

@cindex Alliant
On the Alliant, the system's own convention for returning structures
and unions is unusual, and is not compatible with GNU CC no matter
what options are used.

@cindex RT PC
@cindex IBM RT PC
On the IBM RT PC, the MetaWare HighC compiler (hc) uses yet another
convention for structure and union returning.  Use
@samp{-mhc-struct-return} to tell GNU CC to use a convention compatible
with it.

@cindex Vax calling convention
@cindex Ultrix calling convention
On Ultrix, the Fortran compiler expects registers 2 through 5 to be saved
by function calls.  However, the C compiler uses conventions compatible
with BSD Unix: registers 2 through 5 may be clobbered by function calls.

GNU CC uses the same convention as the Ultrix C compiler.  You can use
these options to produce code compatible with the Fortran compiler:

-fcall-saved-r2 -fcall-saved-r3 -fcall-saved-r4 -fcall-saved-r5
@end smallexample

On the WE32k, you may find that programs compiled with GNU CC do not
work with the standard shared C ilbrary.  You may need to link with
the ordinary C compiler.  If you do so, you must specify the following

-L/usr/local/lib/gcc-lib/we32k-att-sysv/2.3 -lgcc -lc_s
@end smallexample

The first specifies where to find the library @file{libgcc.a}
specified with the @samp{-lgcc} option.

GNU CC does linking by invoking @code{ld}, just as @code{cc} does, and
there is no reason why it @emph{should} matter which compilation program
you use to invoke @code{ld}.  If someone tracks this problem down,
it can probably be fixed easily.
@end itemize

@node Incompatibilities
@section Incompatibilities of GNU CC
@cindex incompatibilities of GNU CC

There are several noteworthy incompatibilities between GNU C and most
existing (non-ANSI) versions of C.  The @samp{-traditional} option
eliminates many of these incompatibilities, @emph{but not all}, by
telling GNU C to behave like the other C compilers.

@itemize @bullet
@cindex string constants
@cindex read-only strings
@cindex shared strings
GNU CC normally makes string constants read-only.  If several
identical-looking string constants are used, GNU CC stores only one
copy of the string.

@cindex @code{mktemp}, and constant strings
One consequence is that you cannot call @code{mktemp} with a string
constant argument.  The function @code{mktemp} always alters the
string its argument points to.

@cindex @code{sscanf}, and constant strings
@cindex @code{fscanf}, and constant strings
@cindex @code{scanf}, and constant strings
Another consequence is that @code{sscanf} does not work on some systems
when passed a string constant as its format control string or input.
This is because @code{sscanf} incorrectly tries to write into the string
constant.  Likewise @code{fscanf} and @code{scanf}.

The best solution to these problems is to change the program to use
@code{char}-array variables with initialization strings for these
purposes instead of string constants.  But if this is not possible,
you can use the @samp{-fwritable-strings} flag, which directs GNU CC
to handle string constants the same way most C compilers do.
@samp{-traditional} also has this effect, among others.

@code{-2147483648} is positive.

This is because 2147483648 cannot fit in the type @code{int}, so
(following the ANSI C rules) its data type is @code{unsigned long int}.
Negating this value yields 2147483648 again.

GNU CC does not substitute macro arguments when they appear inside of
string constants.  For example, the following macro in GNU CC

#define foo(a) "a"
@end example

will produce output @code{"a"} regardless of what the argument @var{a} is.

The @samp{-traditional} option directs GNU CC to handle such cases
(among others) in the old-fashioned (non-ANSI) fashion.

@cindex @code{setjmp} incompatibilities
@cindex @code{longjmp} incompatibilities
When you use @code{setjmp} and @code{longjmp}, the only automatic
variables guaranteed to remain valid are those declared
@code{volatile}.  This is a consequence of automatic register
allocation.  Consider this function:

jmp_buf j;

foo ()
  int a, b;

  a = fun1 ();
  if (setjmp (j))
    return a;

  a = fun2 ();
  /* @r{@code{longjmp (j)} may occur in @code{fun3}.} */
  return a + fun3 ();
@end example

Here @code{a} may or may not be restored to its first value when the
@code{longjmp} occurs.  If @code{a} is allocated in a register, then
its first value is restored; otherwise, it keeps the last value stored
in it.

If you use the @samp{-W} option with the @samp{-O} option, you will
get a warning when GNU CC thinks such a problem might be possible.

The @samp{-traditional} option directs GNU C to put variables in
the stack by default, rather than in registers, in functions that
call @code{setjmp}.  This results in the behavior found in
traditional C compilers.

Programs that use preprocessor directives in the middle of macro
arguments do not work with GNU CC.  For example, a program like this
will not work:

foobar (
#define luser
@end example

ANSI C does not permit such a construct.  It would make sense to support
it when @samp{-traditional} is used, but it is too much work to

@cindex external declaration scope
@cindex scope of external declarations
@cindex declaration scope
Declarations of external variables and functions within a block apply
only to the block containing the declaration.  In other words, they
have the same scope as any other declaration in the same place.

In some other C compilers, a @code{extern} declaration affects all the
rest of the file even if it happens within a block.

The @samp{-traditional} option directs GNU C to treat all @code{extern}
declarations as global, like traditional compilers.

In traditional C, you can combine @code{long}, etc., with a typedef name,
as shown here:

typedef int foo;
typedef long foo bar;
@end example

In ANSI C, this is not allowed: @code{long} and other type modifiers
require an explicit @code{int}.  Because this criterion is expressed
by Bison grammar rules rather than C code, the @samp{-traditional}
flag cannot alter it.

@cindex typedef names as function parameters
PCC allows typedef names to be used as function parameters.  The
difficulty described immediately above applies here too.

@cindex whitespace
PCC allows whitespace in the middle of compound assignment operators
such as @samp{+=}.  GNU CC, following the ANSI standard, does not
allow this.  The difficulty described immediately above applies here

@cindex apostrophes
@cindex '
GNU CC complains about unterminated character constants inside of
preprocessor conditionals that fail.  Some programs have English
comments enclosed in conditionals that are guaranteed to fail; if these
comments contain apostrophes, GNU CC will probably report an error.  For
example, this code would produce an error:

#if 0
You can't expect this to work.
@end example

The best solution to such a problem is to put the text into an actual
C comment delimited by @samp{/*@dots{}*/}.  However,
@samp{-traditional} suppresses these error messages.

Many user programs contain the declaration @samp{long time ();}.  In the
past, the system header files on many systems did not actually declare
@code{time}, so it did not matter what type your program declared it to
return.  But in systems with ANSI C headers, @code{time} is declared to
return @code{time_t}, and if that is not the same as @code{long}, then
@samp{long time ();} is erroneous.

The solution is to change your program to use @code{time_t} as the return
type of @code{time}.

@cindex @code{float} as function value type
When compiling functions that return @code{float}, PCC converts it to
a double.  GNU CC actually returns a @code{float}.  If you are concerned
with PCC compatibility, you should declare your functions to return
@code{double}; you might as well say what you mean.

@cindex structures
@cindex unions
When compiling functions that return structures or unions, GNU CC
output code normally uses a method different from that used on most
versions of Unix.  As a result, code compiled with GNU CC cannot call
a structure-returning function compiled with PCC, and vice versa.

The method used by GNU CC is as follows: a structure or union which is
1, 2, 4 or 8 bytes long is returned like a scalar.  A structure or union
with any other size is stored into an address supplied by the caller
(usually in a special, fixed register, but on some machines it is passed
on the stack).  The machine-description macros @code{STRUCT_VALUE} and
@code{STRUCT_INCOMING_VALUE} tell GNU CC where to pass this address.

By contrast, PCC on most target machines returns structures and unions
of any size by copying the data into an area of static storage, and then
returning the address of that storage as if it were a pointer value.
The caller must copy the data from that memory area to the place where
the value is wanted.  GNU CC does not use this method because it is
slower and nonreentrant.

On some newer machines, PCC uses a reentrant convention for all
structure and union returning.  GNU CC on most of these machines uses a
compatible convention when returning structures and unions in memory,
but still returns small structures and unions in registers.

You can tell GNU CC to use a compatible convention for all structure and
union returning with the option @samp{-fpcc-struct-return}.
@end itemize

@node Disappointments
@section Disappointments and Misunderstandings

These problems are perhaps regrettable, but we don't know any practical
way around them.

@itemize @bullet
Certain local variables aren't recognized by debuggers when you compile
with optimization.

This occurs because sometimes GNU CC optimizes the variable out of
existence.  There is no way to tell the debugger how to compute the
value such a variable ``would have had'', and it is not clear that would
be desirable anyway.  So GNU CC simply does not mention the eliminated
variable when it writes debugging information.

You have to expect a certain amount of disagreement between the
executable and your source code, when you use optimization.

@cindex conflicting types
@cindex scope of declaration
Users often think it is a bug when GNU CC reports an error for code
like this:

int foo (struct mumble *);

struct mumble @{ @dots{} @};

int foo (struct mumble *x)
@{ @dots{} @}
@end example

This code really is erroneous, because the scope of @code{struct
mumble} in the prototype is limited to the argument list containing it.
It does not refer to the @code{struct mumble} defined with file scope
immediately below---they are two unrelated types with similar names in
different scopes.

But in the definition of @code{foo}, the file-scope type is used
because that is available to be inherited.  Thus, the definition and
the prototype do not match, and you get an error.

This behavior may seem silly, but it's what the ANSI standard specifies.
It is easy enough for you to make your code work by moving the
definition of @code{struct mumble} above the prototype.  It's not worth
being incompatible with ANSI C just to avoid an error for the example
shown above.

Accesses to bitfields even in volatile objects works by accessing larger
objects, such as a byte or a word.  You cannot rely on what size of
object is accessed in order to read or write the bitfield; it may even
vary for a given bitfield according to the precise usage.

If you care about controlling the amount of memory that is accessed, use
volatile but do not use bitfields.

On 68000 systems, you can get paradoxical results if you test the
precise values of floating point numbers.  For example, you can find
that a floating point value which is not a NaN is not equal to itself.
This results from the fact that the the floating point registers hold a
few more bits of precision than fit in a @code{double} in memory.
Compiled code moves values between memory and floating point registers
at its convenience, and moving them into memory truncates them.

You can partially avoid this problem by using the option
@samp{-ffloat-store} (@pxref{Optimize Options}).

On the MIPS, variable argument functions using @file{varargs.h}
cannot have a floating point value for the first argument.  The
reason for this is that in the absence of a prototype in scope,
if the first argument is a floating point, it is passed in a
floating point register, rather than an integer resgister.

If the code is rewritten to use the ANSI standard @file{stdarg.h}
method of variable arguments, and the prototype is in scope at
the time of the call, everything will work fine.
@end itemize

@node Protoize Caveats
@section Caveats of using @code{protoize}

The conversion programs @code{protoize} and @code{unprotoize} can
sometimes change a source file in a way that won't work unless you
rearrange it.

@itemize @bullet
@code{protoize} can insert references to a type name or type tag before
the definition, or in a file where they are not defined.

If this happens, compiler error messages should show you where the new
references are, so fixing the file by hand is straightforward.

There are some C constructs which @code{protoize} cannot figure out.
For example, it can't determine argument types for declaring a
pointer-to-function variable; this you must do by hand.  @code{protoize}
inserts a comment containing @samp{???} each time it finds such a
variable; so you can find all such variables by searching for this
string.  ANSI C does not require declaring the argument types of
pointer-to-function types.

Using @code{unprotoize} can easily introduce bugs.  If the program
relied on prototypes to bring about conversion of arguments, these
conversions will not take place in the program without prototypes.
One case in which you can be sure @code{unprotoize} is safe is when
you are removing prototypes that were made with @code{protoize}; if
the program worked before without any prototypes, it will work again
without them.

You can find all the places where this problem might occur by compiling
the program with the @samp{-Wconversion} option.  It prints a warning
whenever an argument is converted.

Both conversion programs can be confused if there are macro calls in and
around the text to be converted.  In other words, the standard syntax
for a declaration or definition must not result from expanding a macro.
This problem is inherent in the design of C and cannot be fixed.  If
only a few functions have confusing macro calls, you can easily convert
them manually.

@code{protoize} cannot get the argument types for a function whose
definition was not actually compiled due to preprocessor conditionals.
When this happens, @code{protoize} changes nothing in regard to such 
a function.  @code{protoize} tries to detect such instances and warn
about them.

You can generally work around this problem by using @code{protoize} step
by step, each time specifying a different set of @samp{-D} options for
compilation, until all of the functions have been converted.  There is
no automatic way to verify that you have got them all, however.

Confusion may result if there is an occasion to convert a function
declaration or definition in a region of source code where there is more
than one formal parameter list present.  Thus, attempts to convert code
containing multiple (conditionally compiled) versions of a single
function header (in the same vicinity) may not produce the desired (or
expected) results.

If you plan on converting source files which contain such code, it is
recommended that you first make sure that each conditionally compiled
region of source code which contains an alternative function header also
contains at least one additional follower token (past the final right
parenthesis of the function header).  This should circumvent the

@code{unprotoize} can become confused when trying to convert a function
definition or declaration which contains a declaration for a
pointer-to-function formal argument which has the same name as the
function being defined or declared.  We recommand you avoid such choices
of formal parameter names.

You might also want to correct some of the indentation by hand and break
long lines.  (The conversion programs don't write lines longer than
eighty characters in any case.)
@end itemize

@node Non-bugs
@section Certain Changes We Don't Want to Make

This section lists changes that people frequently request, but which
we do not make because we think GNU CC is better without them.

@itemize @bullet
Checking the number and type of arguments to a function which has an
old-fashioned definition and no prototype.

Such a feature would work only occasionally---only for calls that appear
in the same file as the called function, following the definition.  The
only way to check all calls reliably is to add a prototype for the
function.  But adding a prototype eliminates the motivation for this
feature.  So the feature is not worthwhile.

Warning about using an expression whose type is signed as a shift count.

Shift count operands are probably signed more often than unsigned.
Warning about this would cause far more annoyance than good.

Warning about assigning a signed value to an unsigned variable.

Such assignments must be very common; warning about them would cause
more annoyance than good.

Warning about unreachable code.

It's very common to have unreachable code in machine-generated
programs.  For example, this happens normally in some files of GNU C

Warning when a non-void function value is ignored.

Coming as I do from a Lisp background, I balk at the idea that there is
something dangerous about discarding a value.  There are functions that
return values which some callers may find useful; it makes no sense to
clutter the program with a cast to @code{void} whenever the value isn't

Assuming (for optimization) that the address of an external symbol is
never zero.

This assumption is false on certain systems when @samp{#pragma weak} is

Making @samp{-fshort-enums} the default.

This would cause storage layout to be incompatible with most other C
compilers.  And it doesn't seem very important, given that you can get
the same result in other ways.  The case where it matters most is when
the enumeration-valued object is inside a structure, and in that case
you can specify a field width explicitly.

Making bitfields unsigned by default on particular machines where ``the
ABI standard'' says to do so.

The ANSI C standard leaves it up to the implementation whether a bitfield
declared plain @code{int} is signed or not.  This in effect creates two
alternative dialects of C.

The GNU C compiler supports both dialects; you can specify the dialect
you want with the option @samp{-fsigned-bitfields} or
@samp{-funsigned-bitfields}.  However, this leaves open the question
of which dialect to use by default.

Currently, the preferred dialect makes plain bitfields signed, because
this is simplest.  Since @code{int} is the same as @code{signed int} in
every other context, it is cleanest for them to be the same in bitfields
as well.

Some computer manufacturers have published Application Binary Interface
standards which specify that plain bitfields should be unsigned.  It is
a mistake, however, to say anything about this issue in an ABI.  This is
because the handling of plain bitfields distinguishes two dialects of C.
Both dialects are meaningful on every type of machine.  Whether a
particular object file was compiled using signed bitfields or unsigned
is of no concern to other object files, even if they access the same
bitfields in the same data structures.

A given program is written in one or the other of these two dialects.
The program stands a chance to work on most any machine if it is
compiled with the proper dialect.  It is unlikely to work at all if
compiled with the wrong dialect.

Many users appreciate the GNU C compiler because it provides an
environment that is uniform across machines.  These users would be
inconvenienced if the compiler treated plain bitfields differently on
certain machines.

Occasionally users write programs intended only for a particular machine
type.  On these occasions, the users would benefit if the GNU C compiler
were to support by default the same dialect as the other compilers on
that machine.  But such applications are rare.  And users writing a
program to run on more than one type of machine cannot possibly benefit
from this kind of compatibility.

This is why GNU CC does and will treat plain bitfields in the same
fashion on all types of machines (by default).

There are some arguments for making bitfields unsigned by default on all
machines.  If, for example, this becomes a universal de facto standard,
it would make sense for GNU CC to go along with it.  This is something
to be considered in the future.

(Of course, users strongly concerned about portability should indicate
explicitly in each bitfield whether it is signed or not.  In this way,
they write programs which have the same meaning in both C dialects.)

Undefining @code{__STDC__} when @samp{-ansi} is not used.

Currently, GNU CC defines @code{__STDC__} as long as you don't use
@samp{-traditional}.  This provides good results in practice.

Programmers normally use conditionals on @code{__STDC__} to ask whether
it is safe to use certain features of ANSI C, such as function
prototypes or ANSI token concatenation.  Since plain @samp{gcc} supports
all the features of ANSI C, the correct answer to these questions is

Some users try to use @code{__STDC__} to check for the availability of
certain library facilities.  This is actually incorrect usage in an ANSI
C program, because the ANSI C standard says that a conforming
freestanding implementation should define @code{__STDC__} even though it
does not have the library facilities.  @samp{gcc -ansi -pedantic} is a
conforming freestanding implementation, and it is therefore required to
define @code{__STDC__}, even though it does not come with an ANSI C

Sometimes people say that defining @code{__STDC__} in a compiler that
does not completely conform to the ANSI C standard somehow violates the
standard.  This is illogical.  The standard is a standard for compilers
that claim to support ANSI C, such as @samp{gcc -ansi}---not for other
compilers such as plain @samp{gcc}.  Whatever the ANSI C standard says
is relevant to the design of plain @samp{gcc} without @samp{-ansi} only
for pragmatic reasons, not as a requirement.

Undefining @code{__STDC__} in C++.

Programs written to compile with C++-to-C translators get the
value of @code{__STDC__} that goes with the C compiler that is
subsequently used.  These programs must test @code{__STDC__}
to determine what kind of C preprocessor that compiler uses:
whether they should concatenate tokens in the ANSI C fashion
or in the traditional fashion.

These programs work properly with GNU C++ if @code{__STDC__} is defined.
They would not work otherwise.

In addition, many header files are written to provide prototypes in ANSI
C but not in traditional C.  Many of these header files can work without
change in C++ provided @code{__STDC__} is defined.  If @code{__STDC__}
is not defined, they will all fail, and will all need to be changed to
test explicitly for C++ as well.

Deleting ``empty'' loops.

GNU CC does not delete ``empty'' loops because the most likely reason
you would put one in a program is to have a delay.  Deleting them will
not make real programs run any faster, so it would be pointless.

It would be different if optimization of a nonempty loop could produce
an empty one.  But this generally can't happen.
@end itemize

@node Bugs
@chapter Reporting Bugs
@cindex bugs
@cindex reporting bugs

Your bug reports play an essential role in making GNU CC reliable.

When you encounter a problem, the first thing to do is to see if it is
already known.  @xref{Trouble}.  If it isn't known, then you should
report the problem.

Reporting a bug may help you by bringing a solution to your problem, or
it may not.  (If it does not, look in the service directory; see
@ref{Service}.)  In any case, the principal function of a bug report is
to help the entire community by making the next version of GNU CC work
better.  Bug reports are your contribution to the maintenance of GNU CC.

In order for a bug report to serve its purpose, you must include the
information that makes for fixing the bug.

* Criteria:  Bug Criteria.   Have you really found a bug?
* Where: Bug Lists.	     Where to send your bug report.
* Reporting: Bug Reporting.  How to report a bug effectively.
* Patches: Sending Patches.  How to send a patch for GNU CC.
* Known: Trouble.            Known problems.
* Help: Service.             Where to ask for help.
@end menu

@node Bug Criteria
@section Have You Found a Bug?
@cindex bug criteria

If you are not sure whether you have found a bug, here are some guidelines:

@itemize @bullet
@cindex fatal signal
@cindex core dump
If the compiler gets a fatal signal, for any input whatever, that is a
compiler bug.  Reliable compilers never crash.

@cindex invalid assembly code
@cindex assembly code, invalid
If the compiler produces invalid assembly code, for any input whatever
(except an @code{asm} statement), that is a compiler bug, unless the
compiler reports errors (not just warnings) which would ordinarily
prevent the assembler from being run.

@cindex undefined behavior
@cindex undefined function value
@cindex increment operators
If the compiler produces valid assembly code that does not correctly
execute the input source code, that is a compiler bug.

However, you must double-check to make sure, because you may have run
into an incompatibility between GNU C and traditional C
(@pxref{Incompatibilities}).  These incompatibilities might be considered
bugs, but they are inescapable consequences of valuable features.

Or you may have a program whose behavior is undefined, which happened
by chance to give the desired results with another C or C++ compiler.

For example, in many nonoptimizing compilers, you can write @samp{x;}
at the end of a function instead of @samp{return x;}, with the same
results.  But the value of the function is undefined if @code{return}
is omitted; it is not a bug when GNU CC produces different results.

Problems often result from expressions with two increment operators,
as in @code{f (*p++, *p++)}.  Your previous compiler might have
interpreted that expression the way you intended; GNU CC might
interpret it another way.  Neither compiler is wrong.  The bug is
in your code.

After you have localized the error to a single source line, it should
be easy to check for these things.  If your program is correct and
well defined, you have found a compiler bug.

If the compiler produces an error message for valid input, that is a
compiler bug.

@cindex invalid input
If the compiler does not produce an error message for invalid input,
that is a compiler bug.  However, you should note that your idea of
``invalid input'' might be my idea of ``an extension'' or ``support
for traditional practice''.

If you are an experienced user of C or C++ compilers, your suggestions
for improvement of GNU CC or GNU C++ are welcome in any case.
@end itemize

@node Bug Lists
@section Where to Report Bugs
@cindex bug report mailing lists

Send bug reports for GNU C to one of these addresses:

@end example

Send bug reports for GNU C++ to one of these addresses:

@end example

@strong{Do not send bug reports to @samp{help-gcc}, or to the newsgroup
@samp{gnu.gcc.help}.}  Most users of GNU CC do not want to receive bug
reports.  Those that do, have asked to be on @samp{bug-gcc} and/or

The mailing lists @samp{bug-gcc} and @samp{bug-g++} both have newsgroups
which serve as repeaters: @samp{gnu.gcc.bug} and @samp{gnu.g++.bug}.
Each mailing list and its newsgroup carry exactly the same messages.

Often people think of posting bug reports to the newsgroup instead of
mailing them.  This appears to work, but it has one problem which can be
crucial: a newsgroup posting does not contain a mail path back to the
sender.  Thus, if maintainers need more information, they may be unable
to reach you.  For this reason, you should always send bug reports by
mail to the proper mailing list.

As a last resort, send bug reports on paper to:

GNU Compiler Bugs
Free Software Foundation
675 Mass Ave
Cambridge, MA 02139
@end example

@node Bug Reporting
@section How to Report Bugs
@cindex compiler bugs, reporting

The fundamental principle of reporting bugs usefully is this:
@strong{report all the facts}.  If you are not sure whether to state a
fact or leave it out, state it!

Often people omit facts because they think they know what causes the
problem and they conclude that some details don't matter.  Thus, you might
assume that the name of the variable you use in an example does not matter.
Well, probably it doesn't, but one cannot be sure.  Perhaps the bug is a
stray memory reference which happens to fetch from the location where that
name is stored in memory; perhaps, if the name were different, the contents
of that location would fool the compiler into doing the right thing despite
the bug.  Play it safe and give a specific, complete example.  That is the
easiest thing for you to do, and the most helpful.

Keep in mind that the purpose of a bug report is to enable someone to
fix the bug if it is not known.  It isn't very important what happens if
the bug is already known.  Therefore, always write your bug reports on
the assumption that the bug is not known.

Sometimes people give a few sketchy facts and ask, ``Does this ring a
bell?''  This cannot help us fix a bug, so it is basically useless.  We
respond by asking for enough details to enable us to investigate.
You might as well expedite matters by sending them to begin with.

Try to make your bug report self-contained.  If we have to ask you for
more information, it is best if you include all the previous information
in your response, as well as the information that was missing.

To enable someone to investigate the bug, you should include all these

@itemize @bullet
The version of GNU CC.  You can get this by running it with the
@samp{-v} option.

Without this, we won't know whether there is any point in looking for
the bug in the current version of GNU CC.

A complete input file that will reproduce the bug.  If the bug is in the
C preprocessor, send a source file and any header files that it
requires.  If the bug is in the compiler proper (@file{cc1}), run your
source file through the C preprocessor by doing @samp{gcc -E
@var{sourcefile} > @var{outfile}}, then include the contents of
@var{outfile} in the bug report.  (When you do this, use the same
@samp{-I}, @samp{-D} or @samp{-U} options that you used in actual

A single statement is not enough of an example.  In order to compile it,
it must be embedded in a complete file of compiler input; and the bug
might depend on the details of how this is done.

Without a real example one can compile, all anyone can do about your bug
report is wish you luck.  It would be futile to try to guess how to
provoke the bug.  For example, bugs in register allocation and reloading
frequently depend on every little detail of the function they happen in.

Even if the input file that fails comes from a GNU program, you should
still send the complete test case.  Don't ask the GNU CC maintainers to
do the extra work of obtaining the program in question---they are all
overworked as it is.  Also, the problem may depend on what is in the
header files on your system; it is unreliable for the GNU CC maintainers
to try the problem with the header files available to them.  By sending
CPP output, you can eliminate this source of uncertainty.

The command arguments you gave GNU CC or GNU C++ to compile that example
and observe the bug.  For example, did you use @samp{-O}?  To guarantee
you won't omit something important, list all the options.

If we were to try to guess the arguments, we would probably guess wrong
and then we would not encounter the bug.

The type of machine you are using, and the operating system name and
version number.

The operands you gave to the @code{configure} command when you installed
the compiler.

A complete list of any modifications you have made to the compiler
source.  (We don't promise to investigate the bug unless it happens in
an unmodified compiler.  But if you've made modifications and don't tell
us, then you are sending us on a wild goose chase.)

Be precise about these changes.  A description in English is not
enough---send a context diff for them.

Adding files of your own (such as a machine description for a machine we
don't support) is a modification of the compiler source.

Details of any other deviations from the standard procedure for installing

A description of what behavior you observe that you believe is
incorrect.  For example, ``The compiler gets a fatal signal,'' or,
``The assembler instruction at line 208 in the output is incorrect.''

Of course, if the bug is that the compiler gets a fatal signal, then one
can't miss it.  But if the bug is incorrect output, the maintainer might
not notice unless it is glaringly wrong.  None of us has time to study
all the assembler code from a 50-line C program just on the chance that
one instruction might be wrong.  We need @emph{you} to do this part!

Even if the problem you experience is a fatal signal, you should still
say so explicitly.  Suppose something strange is going on, such as, your
copy of the compiler is out of synch, or you have encountered a bug in
the C library on your system.  (This has happened!)  Your copy might
crash and the copy here would not.  If you @i{said} to expect a crash,
then when the compiler here fails to crash, we would know that the bug
was not happening.  If you don't say to expect a crash, then we would
not know whether the bug was happening.  We would not be able to draw
any conclusion from our observations.

If the problem is a diagnostic when compiling GNU CC with some other
compiler, say whether it is a warning or an error.

Often the observed symptom is incorrect output when your program is run.
Sad to say, this is not enough information unless the program is short
and simple.  None of us has time to study a large program to figure out
how it would work if compiled correctly, much less which line of it was
compiled wrong.  So you will have to do that.  Tell us which source line
it is, and what incorrect result happens when that line is executed.  A
person who understands the program can find this as easily as finding a
bug in the program itself.

If you send examples of assembler code output from GNU CC or GNU C++,
please use @samp{-g} when you make them.  The debugging information
includes source line numbers which are essential for correlating the
output with the input.

If you wish to mention something in the GNU CC source, refer to it by
context, not by line number.

The line numbers in the development sources don't match those in your
sources.  Your line numbers would convey no useful information to the

Additional information from a debugger might enable someone to find a
problem on a machine which he does not have available.  However, you
need to think when you collect this information if you want it to have
any chance of being useful.

@cindex backtrace for bug reports
For example, many people send just a backtrace, but that is never
useful by itself.  A simple backtrace with arguments conveys little
about GNU CC because the compiler is largely data-driven; the same
functions are called over and over for different RTL insns, doing
different things depending on the details of the insn.

Most of the arguments listed in the backtrace are useless because they
are pointers to RTL list structure.  The numeric values of the
pointers, which the debugger prints in the backtrace, have no
significance whatever; all that matters is the contents of the objects
they point to (and most of the contents are other such pointers).

In addition, most compiler passes consist of one or more loops that
scan the RTL insn sequence.  The most vital piece of information about
such a loop---which insn it has reached---is usually in a local variable,
not in an argument.

@findex debug_rtx
What you need to provide in addition to a backtrace are the values of
the local variables for several stack frames up.  When a local
variable or an argument is an RTX, first print its value and then use
the GDB command @code{pr} to print the RTL expression that it points
to.  (If GDB doesn't run on your machine, use your debugger to call
the function @code{debug_rtx} with the RTX as an argument.)  In
general, whenever a variable is a pointer, its value is no use
without the data it points to.
@end itemize

Here are some things that are not necessary:

@itemize @bullet
A description of the envelope of the bug.

Often people who encounter a bug spend a lot of time investigating
which changes to the input file will make the bug go away and which
changes will not affect it.

This is often time consuming and not very useful, because the way we
will find the bug is by running a single example under the debugger with
breakpoints, not by pure deduction from a series of examples.  You might
as well save your time for something else.

Of course, if you can find a simpler example to report @emph{instead} of
the original one, that is a convenience.  Errors in the output will be
easier to spot, running under the debugger will take less time, etc.
Most GNU CC bugs involve just one function, so the most straightforward
way to simplify an example is to delete all the function definitions
except the one where the bug occurs.  Those earlier in the file may be
replaced by external declarations if the crucial function depends on
them.  (Exception: inline functions may affect compilation of functions
defined later in the file.)

However, simplification is not vital; if you don't want to do this,
report the bug anyway and send the entire test case you used.

In particular, some people insert conditionals @samp{#ifdef BUG} around
a statement which, if removed, makes the bug not happen.  These are just
clutter; we won't pay any attention to them anyway.  Besides, you should
send us cpp output, and that can't have conditionals.

A patch for the bug.

A patch for the bug is useful if it is a good one.  But don't omit the
necessary information, such as the test case, on the assumption that a
patch is all we need.  We might see problems with your patch and decide
to fix the problem another way, or we might not understand it at all.

Sometimes with a program as complicated as GNU CC it is very hard to
construct an example that will make the program follow a certain path
through the code.  If you don't send the example, we won't be able to
construct one, so we won't be able to verify that the bug is fixed.

And if we can't understand what bug you are trying to fix, or why your
patch should be an improvement, we won't install it.  A test case will
help us to understand.

@xref{Sending Patches}, for guidelines on how to make it easy for us to
understand and install your patches.

A guess about what the bug is or what it depends on.

Such guesses are usually wrong.  Even I can't guess right about such
things without first using the debugger to find the facts.

A core dump file.

We have no way of examining a core dump for your type of machine
unless we have an identical system---and if we do have one,
we should be able to reproduce the crash ourselves.
@end itemize

@node Sending Patches,, Bug Reporting, Bugs
@section Sending Patches for GNU CC

If you would like to write bug fixes or improvements for the GNU C
compiler, that is very helpful.  When you send your changes, please
follow these guidelines to avoid causing extra work for us in studying
the patches.

If you don't follow these guidelines, your information might still be
useful, but using it will take extra work.  Maintaining GNU C is a lot
of work in the best of circumstances, and we can't keep up unless you do
your best to help.

@itemize @bullet
Send an explanation with your changes of what problem they fix or what
improvement they bring about.  For a bug fix, just include a copy of the
bug report, and explain why the change fixes the bug.

(Referring to a bug report is not as good as including it, because then
we will have to look it up, and we have probably already deleted it if
we've already fixed the bug.)

Always include a proper bug report for the problem you think you have
fixed.  We need to convince ourselves that the change is right before
installing it.  Even if it is right, we might have trouble judging it if
we don't have a way to reproduce the problem.

Include all the comments that are appropriate to help people reading the
source in the future understand why this change was needed.

Don't mix together changes made for different reasons.
Send them @emph{individually}.

If you make two changes for separate reasons, then we might not want to
install them both.  We might want to install just one.  If you send them
all jumbled together in a single set of diffs, we have to do extra work
to disentangle them---to figure out which parts of the change serve
which purpose.  If we don't have time for this, we might have to ignore
your changes entirely.

If you send each change as soon as you have written it, with its own
explanation, then the two changes never get tangled up, and we can
consider each one properly without any extra work to disentangle them.

Ideally, each change you send should be impossible to subdivide into
parts that we might want to consider separately, because each of its
parts gets its motivation from the other parts.

Send each change as soon as that change is finished.  Sometimes people
think they are helping us by accumulating many changes to send them all
together.  As explained above, this is absolutely the worst thing you
could do.

Since you should send each change separately, you might as well send it
right away.  That gives us the option of installing it immediately if it
is important.

Use @samp{diff -c} to make your diffs.  Diffs without context are hard
for us to install reliably.  More than that, they make it hard for us to
study the diffs to decide whether we want to install them.  Unidiff
format is better than contextless diffs, but not as easy to read as
@samp{-c} format.

If you have GNU diff, use @samp{diff -cp}, which shows the name of the
function that each change occurs in.

Write the change log entries for your changes.  We get lots of changes,
and we don't have time to do all the change log writing ourselves.

Read the @file{ChangeLog} file to see what sorts of information to put
in, and to learn the style that we use.  The purpose of the change log
is to show people where to find what was changed.  So you need to be
specific about what functions you changed; in large functions, it's
often helpful to indicate where within the function the change was.

On the other hand, once you have shown people where to find the change,
you need not explain its purpose. Thus, if you add a new function, all
you need to say about it is that it is new.  If you feel that the
purpose needs explaining, it probably does---but the explanation will be
much more useful if you put it in comments in the code.

If you would like your name to appear in the header line for who made
the change, send us the header line.

When you write the fix, keep in mind that we can't install a change that
would break other systems.

People often suggest fixing a problem by changing machine-independent
files such as @file{toplev.c} to do something special that a particular
system needs.  Sometimes it is totally obvious that such changes would
break GNU CC for almost all users.  We can't possibly make a change like
that.  At best it might tell us how to write another patch that would
solve the problem acceptably.

Sometimes people send fixes that @emph{might} be an improvement in
general---but it is hard to be sure of this.  It's hard to install
such changes because we have to study them very carefully.  Of course,
a good explanation of the reasoning by which you concluded the change
was correct can help convince us.

The safest changes are changes to the configuration files for a
particular machine.  These are safe because they can't create new bugs
on other machines.

Please help us keep up with the workload by designing the patch in a
form that is good to install.
@end itemize

@node Service
@chapter How To Get Help with GNU CC

If you need help installing, using or changing GNU CC, there are two
ways to find it:

@itemize @bullet
Send a message to a suitable network mailing list.  First try
@code{bug-gcc@@prep.ai.mit.edu}, and if that brings no response, try

Look in the service directory for someone who might help you for a fee.
The service directory is found in the file named @file{SERVICE} in the
GNU CC distribution.
@end itemize

@node VMS
@chapter Using GNU CC on VMS

* Include Files and VMS::  Where the preprocessor looks for the include files.
* Global Declarations::    How to do globaldef, globalref and globalvalue with
                           GNU CC.
* VMS Misc::		   Misc information.
@end menu

@node Include Files and VMS
@section Include Files and VMS

@cindex include files and VMS
@cindex VMS and include files
@cindex header files and VMS
Due to the differences between the filesystems of Unix and VMS, GNU CC
attempts to translate file names in @samp{#include} into names that VMS
will understand.  The basic strategy is to prepend a prefix to the
specification of the include file, convert the whole filename to a VMS
filename, and then try to open the file.  GNU CC tries various prefixes
one by one until one of them succeeds:

The first prefix is the @samp{GNU_CC_INCLUDE:} logical name: this is
where GNU C header files are traditionally stored.  If you wish to store
header files in non-standard locations, then you can assign the logical
@samp{GNU_CC_INCLUDE} to be a search list, where each element of the
list is suitable for use with a rooted logical.

The next prefix tried is @samp{SYS$SYSROOT:[SYSLIB.]}.  This is where
VAX-C header files are traditionally stored.

If the include file specification by itself is a valid VMS filename, the
preprocessor then uses this name with no prefix in an attempt to open
the include file.

If the file specification is not a valid VMS filename (i.e. does not
contain a device or a directory specifier, and contains a @samp{/}
character), the preprocessor tries to convert it from Unix syntax to 
VMS syntax.

Conversion works like this: the first directory name becomes a device,
and the rest of the directories are converted into VMS-format directory
names.  For example, @file{X11/foobar.h} is translated to
@file{X11:[000000]foobar.h} or @file{X11:foobar.h}, whichever one can be
opened.  This strategy allows you to assign a logical name to point to
the actual location of the header files.

If none of these strategies succeeds, the @samp{#include} fails.
@end enumerate

Include directives of the form:

#include foobar
@end example

are a common source of incompatibility between VAX-C and GNU CC.  VAX-C
treats this much like a standard @code{#include <foobar.h>} directive.
That is incompatible with the ANSI C behavior implemented by GNU CC: to
expand the name @code{foobar} as a macro.  Macro expansion should
eventually yield one of the two standard formats for @code{#include}:

#include "@var{file}"
#include <@var{file}>
@end example

If you have this problem, the best solution is to modify the source to
convert the @code{#include} directives to one of the two standard forms.
That will work with either compiler.  If you want a quick and dirty fix,
define the file names as macros with the proper expansion, like this:

#define stdio <stdio.h>
@end example

This will work, as long as the name doesn't conflict with anything else
in the program.

Another source of incompatibility is that VAX-C assumes that:

#include "foobar"
@end example

is actually asking for the file @file{foobar.h}.  GNU CC does not
make this assumption, and instead takes what you ask for literally;
it tries to read the file @file{foobar}.  The best way to avoid this
problem is to always specify the desired file extension in your include

GNU CC for VMS is distributed with a set of include files that is
sufficient to compile most general purpose programs.  Even though the
GNU CC distribution does not contain header files to define constants
and structures for some VMS system-specific functions, there is no
reason why you cannot use GNU CC with any of these functions.  You first
may have to generate or create header files, either by using the public
domain utility @code{UNSDL} (which can be found on a DECUS tape), or by
extracting the relevant modules from one of the system macro libraries,
and using an editor to construct a C header file.

A @code{#include} file name cannot contain a DECNET node name.  The
preprocessor reports an I/O error if you attempt to use a node name,
whether explicitly, or implicitly via a logical name.

@node Global Declarations
@section Global Declarations and VMS

GNU CC does not provide the @code{globalref}, @code{globaldef} and
@code{globalvalue} keywords of VAX-C.  You can get the same effect with
an obscure feature of GAS, the GNU assembler.  (This requires GAS
version 1.39 or later.)  The following macros allow you to use this
feature in a fairly natural way:

#ifdef __GNUC__
#define GLOBALREF(TYPE,NAME)                      \
  TYPE NAME                                       \
  asm ("_$$PsectAttributes_GLOBALSYMBOL$$" #NAME)
#define GLOBALDEF(TYPE,NAME,VALUE)                \
  TYPE NAME                                       \
  asm ("_$$PsectAttributes_GLOBALSYMBOL$$" #NAME) \
    = VALUE
#define GLOBALVALUEREF(TYPE,NAME)                 \
  const TYPE NAME[1]                              \     
  asm ("_$$PsectAttributes_GLOBALVALUE$$" #NAME)
  const TYPE NAME[1]                              \
  asm ("_$$PsectAttributes_GLOBALVALUE$$" #NAME)  \
    = @{VALUE@}
  globalref TYPE NAME
  globaldef TYPE NAME = VALUE
  globalvalue TYPE NAME = VALUE
  globalvalue TYPE NAME
@end smallexample

(The @code{_$$PsectAttributes_GLOBALSYMBOL} prefix at the start of the
name is removed by the assembler, after it has modified the attributes
of the symbol).  These macros are provided in the VMS binaries
distribution in a header file @file{GNU_HACKS.H}.  An example of the
usage is:

GLOBALREF (int, ijk);
GLOBALDEF (int, jkl, 0);
@end example

The macros @code{GLOBALREF} and @code{GLOBALDEF} cannot be used
straightforwardly for arrays, since there is no way to insert the array
dimension into the declaration at the right place.  However, you can
declare an array with these macros if you first define a typedef for the
array type, like this:

typedef int intvector[10];
GLOBALREF (intvector, foo);
@end example

Array and structure initializers will also break the macros; you can
define the initializer to be a macro of its own, or you can expand the
@code{GLOBALDEF} macro by hand.  You may find a case where you wish to
use the @code{GLOBALDEF} macro with a large array, but you are not
interested in explicitly initializing each element of the array.  In
such cases you can use an initializer like: @code{@{0,@}}, which will
initialize the entire array to @code{0}.

A shortcoming of this implementation is that a variable declared with
@code{GLOBALVALUEREF} or @code{GLOBALVALUEDEF} is always an array.  For
example, the declaration:

@end example

declares the variable @code{ijk} as an array of type @code{int [1]}.
This is done because a globalvalue is actually a constant; its ``value''
is what the linker would normally consider an address.  That is not how
an integer value works in C, but it is how an array works.  So treating
the symbol as an array name gives consistent results---with the
exception that the value seems to have the wrong type.  @strong{Don't
try to access an element of the array.}  It doesn't have any elements.
The array ``address'' may not be the address of actual storage.

The fact that the symbol is an array may lead to warnings where the
variable is used.  Insert type casts to avoid the warnings.  Here is an
example; it takes advantage of the ANSI C feature allowing macros that
expand to use the same name as the macro itself.

GLOBALVALUEREF (int, ss$_normal);
GLOBALVALUEDEF (int, xyzzy,123);
#ifdef __GNUC__
#define ss$_normal ((int) ss$_normal)
#define xyzzy ((int) xyzzy)
@end example

Don't use @code{globaldef} or @code{globalref} with a variable whose
type is an enumeration type; this is not implemented.  Instead, make the
variable an integer, and use a @code{globalvaluedef} for each of the
enumeration values.  An example of this would be:

#ifdef __GNUC__
GLOBALDEF (int, color, 0);
enum globaldef color @{RED, BLUE, GREEN = 3@};
@end example

@node VMS Misc
@section Other VMS Issues

@cindex exit status and VMS
@cindex return value of @code{main}
@cindex @code{main} and the exit status
GNU CC automatically arranges for @code{main} to return 1 by default if
you fail to specify an explicit return value.  This will be interpreted
by VMS as a status code indicating a normal successful completion.
Version 1 of GNU CC did not provide this default.

GNU CC on VMS works only with the GNU assembler, GAS.  You need version
1.37 or later of GAS in order to produce value debugging information for
the VMS debugger.  Use the ordinary VMS linker with the object files
produced by GAS.

@cindex shared VMS run time system
@cindex @file{VAXCRTL}
Under previous versions of GNU CC, the generated code would occasionally
give strange results when linked to the sharable @file{VAXCRTL} library.
Now this should work.

A caveat for use of @code{const} global variables: the @code{const}
modifier must be specified in every external declaration of the variable
in all of the source files that use that variable.  Otherwise the linker
will issue warnings about conflicting attributes for the variable.  Your
program will still work despite the warnings, but the variable will be
placed in writable storage.

@cindex name augmentation
@cindex case sensitivity and VMS
@cindex VMS and case sensitivity
Although the VMS linker does distinguish between upper and lower case
letters in global symbols, most VMS compilers convert all such symbols
into upper case and most run-time library routines also have upper case
names.  To be able to reliably call such routines, GNU CC (by means of
the assembler GAS) converts global symbols into upper case like other
VMS compilers.  However, since the usual practice in C is to distinguish
case, GNU CC (via GAS) tries to preserve usual C behavior by augmenting
each name that is not all lower case.  This means truncating the name
to at most 23 characters and then adding more characters at the end
which encode the case pattern of those 23.   Names which contain at
least one dollar sign are an exception; they are converted directly into
upper case without augmentation.

Name augmentation yields bad results for programs that use precompiled
libraries (such as Xlib) which were generated by another compiler.  You
can use the compiler option @samp{/NOCASE_HACK} to inhibit augmentation;
it makes external C functions and variables case-independent as is usual
on VMS.  Alternatively, you could write all references to the functions
and variables in such libraries using lower case; this will work on VMS,
but is not portable to other systems.  The compiler option @samp{/NAMES}
also provides control over global name handling.

Function and variable names are handled somewhat differently with GNU
C++.  The GNU C++ compiler performs @dfn{name mangling} on function
names, which means that it adds information to the function name to
describe the data types of the arguments that the function takes. One
result of this is that the name of a function can become very long.
Since the VMS linker only recognizes the first 31 characters in a name,
special action is taken to ensure that each function and variable has a
unique name that can be represented in 31 characters.

If the name (plus a name augmentation, if required) is less than 32
characters in length, then no special action is performed. If the name
is longer than 31 characters, the assembler (GAS) will generate a
hash string based upon the function name, truncate the function name to
23 characters, and append the hash string to the truncated name.  If the
@samp{/VERBOSE} compiler option is used, the assembler will print both
the full and truncated names of each symbol that is truncated.

The @samp{/NOCASE_HACK} compiler option should not be used when you are
compiling programs that use libg++. libg++ has several instances of
objects (i.e.  @code{Filebuf} and @code{filebuf}) which become
indistinguishable in a case-insensitive environment.  This leads to
cases where you need to inhibit augmentation selectively (if you were
using libg++ and Xlib in the same program, for example).  There is no
special feature for doing this, but you can get the result by defining a
macro for each mixed case symbol for which you wish to inhibit
augmentation.  The macro should expand into the lower case equivalent of
itself.  For example:

#define StuDlyCapS studlycaps
@end example

These macro definitions can be placed in a header file to minimize the
number of changes to your source code.

@node Portability
@chapter GNU CC and Portability
@cindex portability
@cindex GNU CC and portability

The main goal of GNU CC was to make a good, fast compiler for machines in
the class that the GNU system aims to run on: 32-bit machines that address
8-bit bytes and have several general registers.  Elegance, theoretical
power and simplicity are only secondary.

GNU CC gets most of the information about the target machine from a machine
description which gives an algebraic formula for each of the machine's
instructions.  This is a very clean way to describe the target.  But when
the compiler needs information that is difficult to express in this
fashion, I have not hesitated to define an ad-hoc parameter to the machine
description.  The purpose of portability is to reduce the total work needed
on the compiler; it was not of interest for its own sake.

@cindex endianness
@cindex autoincrement addressing, availability
@findex abort
GNU CC does not contain machine dependent code, but it does contain code
that depends on machine parameters such as endianness (whether the most
significant byte has the highest or lowest address of the bytes in a word)
and the availability of autoincrement addressing.  In the RTL-generation
pass, it is often necessary to have multiple strategies for generating code
for a particular kind of syntax tree, strategies that are usable for different
combinations of parameters.  Often I have not tried to address all possible
cases, but only the common ones or only the ones that I have encountered.
As a result, a new target may require additional strategies.  You will know
if this happens because the compiler will call @code{abort}.  Fortunately,
the new strategies can be added in a machine-independent fashion, and will
affect only the target machines that need them.
@end ifset

@node Interface
@chapter Interfacing to GNU CC Output
@cindex interfacing to GNU CC output
@cindex run-time conventions
@cindex function call conventions
@cindex conventions, run-time

GNU CC is normally configured to use the same function calling convention
normally in use on the target system.  This is done with the
machine-description macros described (@pxref{Target Macros}).

@cindex unions, returning
@cindex structures, returning
@cindex returning structures and unions
However, returning of structure and union values is done differently on
some target machines.  As a result, functions compiled with PCC
returning such types cannot be called from code compiled with GNU CC,
and vice versa.  This does not cause trouble often because few Unix
library routines return structures or unions.

GNU CC code returns structures and unions that are 1, 2, 4 or 8 bytes
long in the same registers used for @code{int} or @code{double} return
values.  (GNU CC typically allocates variables of such types in
registers also.)  Structures and unions of other sizes are returned by
storing them into an address passed by the caller (usually in a
register).  The machine-description macros @code{STRUCT_VALUE} and
@code{STRUCT_INCOMING_VALUE} tell GNU CC where to pass this address.

By contrast, PCC on most target machines returns structures and unions
of any size by copying the data into an area of static storage, and then
returning the address of that storage as if it were a pointer value.
The caller must copy the data from that memory area to the place where
the value is wanted.  This is slower than the method used by GNU CC, and
fails to be reentrant.

On some target machines, such as RISC machines and the 80386, the
standard system convention is to pass to the subroutine the address of
where to return the value.  On these machines, GNU CC has been
configured to be compatible with the standard compiler, when this method
is used.  It may not be compatible for structures of 1, 2, 4 or 8 bytes.

@cindex argument passing
@cindex passing arguments
GNU CC uses the system's standard convention for passing arguments.  On
some machines, the first few arguments are passed in registers; in
others, all are passed on the stack.  It would be possible to use
registers for argument passing on any machine, and this would probably
result in a significant speedup.  But the result would be complete
incompatibility with code that follows the standard convention.  So this
change is practical only if you are switching to GNU CC as the sole C
compiler for the system.  We may implement register argument passing on
certain machines once we have a complete GNU system so that we can
compile the libraries with GNU CC.

On some machines (particularly the Sparc), certain types of arguments
are passed ``by invisible reference''.  This means that the value is
stored in memory, and the address of the memory location is passed to
the subroutine.

@cindex @code{longjmp} and automatic variables
If you use @code{longjmp}, beware of automatic variables.  ANSI C says that
automatic variables that are not declared @code{volatile} have undefined
values after a @code{longjmp}.  And this is all GNU CC promises to do,
because it is very difficult to restore register variables correctly, and
one of GNU CC's features is that it can put variables in registers without
your asking it to.

If you want a variable to be unaltered by @code{longjmp}, and you don't
want to write @code{volatile} because old C compilers don't accept it,
just take the address of the variable.  If a variable's address is ever
taken, even if just to compute it and ignore it, then the variable cannot
go in a register:

  int careful;
@end example

@cindex arithmetic libraries
@cindex math libraries
Code compiled with GNU CC may call certain library routines.  Most of
them handle arithmetic for which there are no instructions.  This
includes multiply and divide on some machines, and floating point
operations on any machine for which floating point support is disabled
with @samp{-msoft-float}.  Some standard parts of the C library, such as
@code{bcopy} or @code{memcpy}, are also called automatically.  The usual
function call interface is used for calling the library routines.

These library routines should be defined in the library @file{libgcc.a},
which GNU CC automatically searches whenever it links a program.  On
machines that have multiply and divide instructions, if hardware
floating point is in use, normally @file{libgcc.a} is not needed, but it
is searched just in case.

Each arithmetic function is defined in @file{libgcc1.c} to use the
corresponding C arithmetic operator.  As long as the file is compiled
with another C compiler, which supports all the C arithmetic operators,
this file will work portably.  However, @file{libgcc1.c} does not work if
compiled with GNU CC, because each arithmetic function would compile
into a call to itself!
@end ifset

@node Passes
@chapter Passes and Files of the Compiler
@cindex passes and files of the compiler
@cindex files and passes of the compiler
@cindex compiler passes and files

@cindex top level of compiler
The overall control structure of the compiler is in @file{toplev.c}.  This
file is responsible for initialization, decoding arguments, opening and
closing files, and sequencing the passes.

@cindex parsing pass
The parsing pass is invoked only once, to parse the entire input.  The RTL
intermediate code for a function is generated as the function is parsed, a
statement at a time.  Each statement is read in as a syntax tree and then
converted to RTL; then the storage for the tree for the statement is
reclaimed.  Storage for types (and the expressions for their sizes),
declarations, and a representation of the binding contours and how they nest,
remain until the function is finished being compiled; these are all needed
to output the debugging information.

@findex rest_of_compilation
@findex rest_of_decl_compilation
Each time the parsing pass reads a complete function definition or
top-level declaration, it calls the function
@code{rest_of_compilation} or @code{rest_of_decl_compilation} in
@file{toplev.c}, which are responsible for all further processing
necessary, ending with output of the assembler language.  All other
compiler passes run, in sequence, within @code{rest_of_compilation}.
When that function returns from compiling a function definition, the
storage used for that function definition's compilation is entirely
freed, unless it is an inline function (@pxref{Inline}).

Here is a list of all the passes of the compiler and their source files.
Also included is a description of where debugging dumps can be requested
with @samp{-d} options.

@itemize @bullet
Parsing.  This pass reads the entire text of a function definition,
constructing partial syntax trees.  This and RTL generation are no longer
truly separate passes (formerly they were), but it is easier to think
of them as separate.

The tree representation does not entirely follow C syntax, because it is
intended to support other languages as well.

Language-specific data type analysis is also done in this pass, and every
tree node that represents an expression has a data type attached.
Variables are represented as declaration nodes.

@cindex constant folding
@cindex arithmetic simplifications
@cindex simplifications, arithmetic
Constant folding and some arithmetic simplifications are also done
during this pass.

The language-independent source files for parsing are
@file{stor-layout.c}, @file{fold-const.c}, and @file{tree.c}.
There are also header files @file{tree.h} and @file{tree.def}
which define the format of the tree representation.@refill

The source files for parsing C are @file{c-parse.y}, @file{c-decl.c},
@file{c-typeck.c}, @file{c-convert.c}, @file{c-lang.c}, and
@file{c-aux-info.c} along with header files @file{c-lex.h}, and

The source files for parsing C++ are @file{cp-parse.y},
@file{cp-cvt.c}, @file{cp-decl.c}, @file{cp-decl2.c},
@file{cp-dem.c}, @file{cp-except.c},@*
@file{cp-expr.c}, @file{cp-init.c}, @file{cp-lex.c},
@file{cp-method.c}, @file{cp-ptree.c},@*
@file{cp-search.c}, @file{cp-tree.c}, @file{cp-type2.c}, and
@file{cp-typeck.c}, along with header files @file{cp-tree.def},
@file{cp-tree.h}, and @file{cp-decl.h}.

The special source files for parsing Objective C are
@file{objc-parse.y}, @file{objc-actions.c}, @file{objc-tree.def}, and
@file{objc-actions.h}.  Certain C-specific files are used for this as

The file @file{c-common.c} is also used for all of the above languages.

@cindex RTL generation
RTL generation.  This is the conversion of syntax tree into RTL code.
It is actually done statement-by-statement during parsing, but for
most purposes it can be thought of as a separate pass.

@cindex target-parameter-dependent code
This is where the bulk of target-parameter-dependent code is found,
since often it is necessary for strategies to apply only when certain
standard kinds of instructions are available.  The purpose of named
instruction patterns is to provide this information to the RTL
generation pass.

@cindex tail recursion optimization
Optimization is done in this pass for @code{if}-conditions that are
comparisons, boolean operations or conditional expressions.  Tail
recursion is detected at this time also.  Decisions are made about how
best to arrange loops and how to output @code{switch} statements.

The source files for RTL generation include @file{stmt.c},
@file{function.c}, @file{expr.c}, @file{calls.c}, @file{explow.c},
@file{expmed.c}, @file{optabs.c} and @file{emit-rtl.c}.  Also, the file
@file{insn-emit.c}, generated from the machine description by the
program @code{genemit}, is used in this pass.  The header file
@file{expr.h} is used for communication within this pass.@refill

@findex genflags
@findex gencodes
The header files @file{insn-flags.h} and @file{insn-codes.h},
generated from the machine description by the programs @code{genflags}
and @code{gencodes}, tell this pass which standard names are available
for use and which patterns correspond to them.@refill

Aside from debugging information output, none of the following passes
refers to the tree structure representation of the function (only
part of which is saved).

@cindex inline, automatic
The decision of whether the function can and should be expanded inline
in its subsequent callers is made at the end of rtl generation.  The
function must meet certain criteria, currently related to the size of
the function and the types and number of parameters it has.  Note that
this function may contain loops, recursive calls to itself
(tail-recursive functions can be inlined!), gotos, in short, all
constructs supported by GNU CC.  The file @file{integrate.c} contains
the code to save a function's rtl for later inlining and to inline that
rtl when the function is called.  The header file @file{integrate.h}
is also used for this purpose.

The option @samp{-dr} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.rtl} to
the input file name.

@cindex jump optimization
@cindex unreachable code
@cindex dead code
Jump optimization.  This pass simplifies jumps to the following
instruction, jumps across jumps, and jumps to jumps.  It deletes
unreferenced labels and unreachable code, except that unreachable code
that contains a loop is not recognized as unreachable in this pass.
(Such loops are deleted later in the basic block analysis.)  It also
converts some code originally written with jumps into sequences of
instructions that directly set values from the results of comparisons,
if the machine has such instructions.

Jump optimization is performed two or three times.  The first time is
immediately following RTL generation.  The second time is after CSE,
but only if CSE says repeated jump optimization is needed.  The
last time is right before the final pass.  That time, cross-jumping
and deletion of no-op move instructions are done together with the
optimizations described above.

The source file of this pass is @file{jump.c}.

The option @samp{-dj} causes a debugging dump of the RTL code after
this pass is run for the first time.  This dump file's name is made by
appending @samp{.jump} to the input file name.

@cindex register use analysis
Register scan.  This pass finds the first and last use of each
register, as a guide for common subexpression elimination.  Its source
is in @file{regclass.c}.

@cindex jump threading
Jump threading.  This pass detects a condition jump that branches to an
identical or inverse test.  Such jumps can be @samp{threaded} through
the second conditional test.  The source code for this pass is in
@file{jump.c}.  This optimization is only performed if
@samp{-fthread-jumps} is enabled.

@cindex common subexpression elimination
@cindex constant propagation
Common subexpression elimination.  This pass also does constant
propagation.  Its source file is @file{cse.c}.  If constant
propagation causes conditional jumps to become unconditional or to
become no-ops, jump optimization is run again when CSE is finished.

The option @samp{-ds} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.cse} to
the input file name.

@cindex loop optimization
@cindex code motion
@cindex strength-reduction
Loop optimization.  This pass moves constant expressions out of loops,
and optionally does strength-reduction and loop unrolling as well.
Its source files are @file{loop.c} and @file{unroll.c}, plus the header
@file{loop.h} used for communication between them.  Loop unrolling uses
some functions in @file{integrate.c} and the header @file{integrate.h}.

The option @samp{-dL} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.loop} to
the input file name.

If @samp{-frerun-cse-after-loop} was enabled, a second common
subexpression elimination pass is performed after the loop optimization
pass.  Jump threading is also done again at this time if it was specified.

The option @samp{-dt} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.cse2} to
the input file name.

@cindex register allocation, stupid
@cindex stupid register allocation
Stupid register allocation is performed at this point in a
nonoptimizing compilation.  It does a little data flow analysis as
well.  When stupid register allocation is in use, the next pass
executed is the reloading pass; the others in between are skipped.
The source file is @file{stupid.c}.

@cindex data flow analysis
@cindex analysis, data flow
@cindex basic blocks
Data flow analysis (@file{flow.c}).  This pass divides the program
into basic blocks (and in the process deletes unreachable loops); then
it computes which pseudo-registers are live at each point in the
program, and makes the first instruction that uses a value point at
the instruction that computed the value.

@cindex autoincrement/decrement analysis
This pass also deletes computations whose results are never used, and
combines memory references with add or subtract instructions to make
autoincrement or autodecrement addressing.

The option @samp{-df} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.flow} to
the input file name.  If stupid register allocation is in use, this
dump file reflects the full results of such allocation.

@cindex instruction combination
Instruction combination (@file{combine.c}).  This pass attempts to
combine groups of two or three instructions that are related by data
flow into single instructions.  It combines the RTL expressions for
the instructions by substitution, simplifies the result using algebra,
and then attempts to match the result against the machine description.

The option @samp{-dc} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.combine}
to the input file name.

@cindex instruction scheduling
@cindex scheduling, instruction
Instruction scheduling (@file{sched.c}).  This pass looks for
instructions whose output will not be available by the time that it is
used in subsequent instructions.  (Memory loads and floating point
instructions often have this behavior on RISC machines).  It re-orders
instructions within a basic block to try to separate the definition and
use of items that otherwise would cause pipeline stalls.

Instruction scheduling is performed twice.  The first time is immediately
after instruction combination and the second is immediately after reload.

The option @samp{-dS} causes a debugging dump of the RTL code after this
pass is run for the first time.  The dump file's name is made by
appending @samp{.sched} to the input file name.

@cindex register class preference pass
Register class preferencing.  The RTL code is scanned to find out
which register class is best for each pseudo register.  The source
file is @file{regclass.c}.

@cindex register allocation
@cindex local register allocation
Local register allocation (@file{local-alloc.c}).  This pass allocates
hard registers to pseudo registers that are used only within one basic
block.  Because the basic block is linear, it can use fast and
powerful techniques to do a very good job.

The option @samp{-dl} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.lreg} to
the input file name.

@cindex global register allocation
Global register allocation (@file{global.c}).  This pass
allocates hard registers for the remaining pseudo registers (those
whose life spans are not contained in one basic block).

@cindex reloading
Reloading.  This pass renumbers pseudo registers with the hardware
registers numbers they were allocated.  Pseudo registers that did not
get hard registers are replaced with stack slots.  Then it finds
instructions that are invalid because a value has failed to end up in
a register, or has ended up in a register of the wrong kind.  It fixes
up these instructions by reloading the problematical values
temporarily into registers.  Additional instructions are generated to
do the copying.

The reload pass also optionally eliminates the frame pointer and inserts
instructions to save and restore call-clobbered registers around calls.

Source files are @file{reload.c} and @file{reload1.c}, plus the header
@file{reload.h} used for communication between them.

The option @samp{-dg} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.greg} to
the input file name.

@cindex instruction scheduling
@cindex scheduling, instruction
Instruction scheduling is repeated here to try to avoid pipeline stalls
due to memory loads generated for spilled pseudo registers.

The option @samp{-dR} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.sched2}
to the input file name.

@cindex cross-jumping
@cindex no-op move instructions
Jump optimization is repeated, this time including cross-jumping
and deletion of no-op move instructions.

The option @samp{-dJ} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.jump2}
to the input file name.

@cindex delayed branch scheduling
@cindex scheduling, delayed branch
Delayed branch scheduling.  This optional pass attempts to find
instructions that can go into the delay slots of other instructions,
usually jumps and calls.  The source file name is @file{reorg.c}.  

The option @samp{-dd} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.dbr}
to the input file name.

@cindex register-to-stack conversion
Conversion from usage of some hard registers to usage of a register
stack may be done at this point.  Currently, this is supported only
for the floating-point registers of the Intel 80387 coprocessor.   The
source file name is @file{reg-stack.c}.

The options @samp{-dk} causes a debugging dump of the RTL code after
this pass.  This dump file's name is made by appending @samp{.stack}
to the input file name.

@cindex final pass
@cindex peephole optimization
Final.  This pass outputs the assembler code for the function.  It is
also responsible for identifying spurious test and compare
instructions.  Machine-specific peephole optimizations are performed
at the same time.  The function entry and exit sequences are generated
directly as assembler code in this pass; they never exist as RTL.

The source files are @file{final.c} plus @file{insn-output.c}; the
latter is generated automatically from the machine description by the
tool @file{genoutput}.  The header file @file{conditions.h} is used
for communication between these files.

@cindex debugging information generation
Debugging information output.  This is run after final because it must
output the stack slot offsets for pseudo registers that did not get
hard registers.  Source files are @file{dbxout.c} for DBX symbol table
format, @file{sdbout.c} for SDB symbol table format, and
@file{dwarfout.c} for DWARF symbol table format.
@end itemize

Some additional files are used by all or many passes:

@itemize @bullet
Every pass uses @file{machmode.def} and @file{machmode.h} which define
the machine modes.

Several passes use @file{real.h}, which defines the default
representation of floating point constants and how to operate on them.

All the passes that work with RTL use the header files @file{rtl.h}
and @file{rtl.def}, and subroutines in file @file{rtl.c}.  The tools
@code{gen*} also use these files to read and work with the machine
description RTL.

@findex genconfig
Several passes refer to the header file @file{insn-config.h} which
contains a few parameters (C macro definitions) generated
automatically from the machine description RTL by the tool

@cindex instruction recognizer
Several passes use the instruction recognizer, which consists of
@file{recog.c} and @file{recog.h}, plus the files @file{insn-recog.c}
and @file{insn-extract.c} that are generated automatically from the
machine description by the tools @file{genrecog} and

Several passes use the header files @file{regs.h} which defines the
information recorded about pseudo register usage, and @file{basic-block.h}
which defines the information recorded about basic blocks.

@file{hard-reg-set.h} defines the type @code{HARD_REG_SET}, a bit-vector
with a bit for each hard register, and some macros to manipulate it.
This type is just @code{int} if the machine has few enough hard registers;
otherwise it is an array of @code{int} and some of the macros expand
into loops.

Several passes use instruction attributes.  A definition of the
attributes defined for a particular machine is in file
@file{insn-attr.h}, which is generated from the machine description by
the program @file{genattr}.  The file @file{insn-attrtab.c} contains
subroutines to obtain the attribute values for insns.  It is generated
from the machine description by the program @file{genattrtab}.@refill
@end itemize
@end ifset

@include rtl.texi
@include md.texi
@include tm.texi

@node Config
@chapter The Configuration File
@cindex configuration file
@cindex @file{xm-@var{machine}.h}

The configuration file @file{xm-@var{machine}.h} contains macro
definitions that describe the machine and system on which the compiler
is running, unlike the definitions in @file{@var{machine}.h}, which
describe the machine for which the compiler is producing output.  Most
of the values in @file{xm-@var{machine}.h} are actually the same on all
machines that GNU CC runs on, so large parts of all configuration files
are identical.  But there are some macros that vary:

@table @code
@findex USG
@item USG
Define this macro if the host system is System V.

@findex VMS
@item VMS
Define this macro if the host system is VMS.

A C expression for the status code to be returned when the compiler
exits after serious errors.

A C expression for the status code to be returned when the compiler
exits without serious errors.

Defined if the host machine stores words of multi-word values in
big-endian order.  (GNU CC does not depend on the host byte ordering
within a word.)

A numeric code distinguishing the floating point format for the host
machine.  See @code{TARGET_FLOAT_FORMAT} in @ref{Storage Layout} for the
alternatives and default.

A C expression for the number of bits in @code{char} on the host

A C expression for the number of bits in @code{short} on the host

A C expression for the number of bits in @code{int} on the host

A C expression for the number of bits in @code{long} on the host

Define this macro to indicate that the host compiler only supports
@code{int} bit fields, rather than other integral types, including
@code{enum}, as do most C compilers.

Define this macro if the host system uses a naming convention for
executable files that involves a common suffix (such as, in some
systems, @samp{.exe}) that must be mentioned explicitly when you run
the program.

A C expression for the size of ordinary obstack chunks.
If you don't define this, a usually-reasonable default is used.

The function used to allocate obstack chunks.
If you don't define this, @code{xmalloc} is used.

The function used to free obstack chunks.
If you don't define this, @code{free} is used.

@findex USE_C_ALLOCA
Define this macro to indicate that the compiler is running with the
@code{alloca} implemented in C.  This version of @code{alloca} can be
found in the file @file{alloca.c}; to use it, you must also alter the
@file{Makefile} variable @code{ALLOCA}.  (This is done automatically
for the systems on which we know it is needed.)

If you do define this macro, you should probably do it as follows:

#ifndef __GNUC__
#define USE_C_ALLOCA
#define alloca __builtin_alloca
@end example

so that when the compiler is compiled with GNU CC it uses the more
efficient built-in @code{alloca} function.

Define this macro to indicate that the host compiler does not properly
handle converting a function value to a pointer-to-function when it is
used in an expression.

@findex vprintf
Define this if the library function @code{vprintf} is available on your

Define this macro to enable support for multibyte characters in the
input to GNU CC.  This requires that the host system support the ANSI C
library functions for converting multibyte characters to wide

@findex putenv
Define this if the library function @code{putenv} is available on your

Define this if your system @emph{does not} provide the variable

Define this to be 1 if you know that the host compiler supports
prototypes, even if it doesn't define __STDC__, or define
it to be 0 if you do not want any prototypes used in compiling
GNU CC.  If @samp{USE_PROTOTYPES} is not defined, it will be
determined automatically whether your compiler supports
prototypes by checking if @samp{__STDC__} is defined.

Define this if you wish suppression of prototypes generated from
the machine description file, but to use other prototypes within
GNU CC.  If @samp{USE_PROTOTYPES} is defined to be 0, or the
host compiler does not support prototypes, this macro has no

Define this if you wish to generate prototypes for the
@code{gen_call} or @code{gen_call_value} functions generated from
the machine description file.  If @samp{USE_PROTOTYPES} is
defined to be 0, or the host compiler does not support
prototypes, or @samp{NO_MD_PROTOTYPES} is defined, this macro has
no effect.  As soon as all of the machine descriptions are
modified to have the appropriate number of arguments, this macro
will be removed.

@vindex sys_siglist
Some systems do provide this variable, but with a different name such
as @code{_sys_siglist}.  On these systems, you can define
@code{sys_siglist} as a macro which expands into the name actually

@findex NO_STAB_H
@item NO_STAB_H
Define this if your system does not have the include file
@file{stab.h}.  If @samp{USG} is defined, @samp{NO_STAB_H} is
@end table

@findex bzero
@findex bcmp
In addition, configuration files for system V define @code{bcopy},
@code{bzero} and @code{bcmp} as aliases.  Some files define @code{alloca}
as a macro when compiled with GNU CC, in order to take advantage of the
benefit of GNU CC's built-in @code{alloca}.

@node Index
@unnumbered Index
@end ifset

@ifclear INTERNALS
@node Index
@unnumbered Index
@end ifclear

@printindex cp