1 \# --------------------------------------------------------------------------
3 \# Copyright 1996-2012 The NASM Authors - All Rights Reserved
4 \# See the file AUTHORS included with the NASM distribution for
5 \# the specific copyright holders.
7 \# Redistribution and use in source and binary forms, with or without
8 \# modification, are permitted provided that the following
11 \# * Redistributions of source code must retain the above copyright
12 \# notice, this list of conditions and the following disclaimer.
13 \# * Redistributions in binary form must reproduce the above
14 \# copyright notice, this list of conditions and the following
15 \# disclaimer in the documentation and/or other materials provided
16 \# with the distribution.
18 \# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
19 \# CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
20 \# INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
21 \# MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
22 \# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
23 \# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 \# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
25 \# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
26 \# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 \# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
28 \# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
29 \# OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
30 \# EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
32 \# --------------------------------------------------------------------------
34 \# Source code to NASM documentation
36 \M{category}{Programming}
37 \M{title}{NASM - The Netwide Assembler}
39 \M{author}{The NASM Development Team}
40 \M{copyright_tail}{-- All Rights Reserved}
41 \M{license}{This document is redistributable under the license given in the file "LICENSE" distributed in the NASM archive.}
42 \M{summary}{This file documents NASM, the Netwide Assembler: an assembler targetting the Intel x86 series of processors, with portable source.}
45 \M{infotitle}{The Netwide Assembler for x86}
46 \M{epslogo}{nasmlogo.eps}
53 \IR{-MD} \c{-MD} option
54 \IR{-MF} \c{-MF} option
55 \IR{-MG} \c{-MG} option
56 \IR{-MP} \c{-MP} option
57 \IR{-MQ} \c{-MQ} option
58 \IR{-MT} \c{-MT} option
79 \IR{!=} \c{!=} operator
80 \IR{$, here} \c{$}, Here token
81 \IR{$, prefix} \c{$}, prefix
84 \IR{%%} \c{%%} operator
85 \IR{%+1} \c{%+1} and \c{%-1} syntax
87 \IR{%0} \c{%0} parameter count
89 \IR{&&} \c{&&} operator
91 \IR{..@} \c{..@} symbol prefix
93 \IR{//} \c{//} operator
95 \IR{<<} \c{<<} operator
96 \IR{<=} \c{<=} operator
97 \IR{<>} \c{<>} operator
99 \IR{==} \c{==} operator
100 \IR{>} \c{>} operator
101 \IR{>=} \c{>=} operator
102 \IR{>>} \c{>>} operator
103 \IR{?} \c{?} MASM syntax
104 \IR{^} \c{^} operator
105 \IR{^^} \c{^^} operator
106 \IR{|} \c{|} operator
107 \IR{||} \c{||} operator
108 \IR{~} \c{~} operator
109 \IR{%$} \c{%$} and \c{%$$} prefixes
111 \IR{+ opaddition} \c{+} operator, binary
112 \IR{+ opunary} \c{+} operator, unary
113 \IR{+ modifier} \c{+} modifier
114 \IR{- opsubtraction} \c{-} operator, binary
115 \IR{- opunary} \c{-} operator, unary
116 \IR{! opunary} \c{!} operator, unary
117 \IR{alignment, in bin sections} alignment, in \c{bin} sections
118 \IR{alignment, in elf sections} alignment, in \c{elf} sections
119 \IR{alignment, in win32 sections} alignment, in \c{win32} sections
120 \IR{alignment, of elf common variables} alignment, of \c{elf} common
122 \IR{alignment, in obj sections} alignment, in \c{obj} sections
123 \IR{a.out, bsd version} \c{a.out}, BSD version
124 \IR{a.out, linux version} \c{a.out}, Linux version
125 \IR{autoconf} Autoconf
127 \IR{bitwise and} bitwise AND
128 \IR{bitwise or} bitwise OR
129 \IR{bitwise xor} bitwise XOR
130 \IR{block ifs} block IFs
131 \IR{borland pascal} Borland, Pascal
132 \IR{borland's win32 compilers} Borland, Win32 compilers
133 \IR{braces, after % sign} braces, after \c{%} sign
135 \IR{c calling convention} C calling convention
136 \IR{c symbol names} C symbol names
137 \IA{critical expressions}{critical expression}
138 \IA{command line}{command-line}
139 \IA{case sensitivity}{case sensitive}
140 \IA{case-sensitive}{case sensitive}
141 \IA{case-insensitive}{case sensitive}
142 \IA{character constants}{character constant}
143 \IR{common object file format} Common Object File Format
144 \IR{common variables, alignment in elf} common variables, alignment
146 \IR{common, elf extensions to} \c{COMMON}, \c{elf} extensions to
147 \IR{common, obj extensions to} \c{COMMON}, \c{obj} extensions to
148 \IR{declaring structure} declaring structures
149 \IR{default-wrt mechanism} default-\c{WRT} mechanism
152 \IR{dll symbols, exporting} DLL symbols, exporting
153 \IR{dll symbols, importing} DLL symbols, importing
155 \IR{dos archive} DOS archive
156 \IR{dos source archive} DOS source archive
157 \IA{effective address}{effective addresses}
158 \IA{effective-address}{effective addresses}
160 \IR{elf, 16-bit code and} ELF, 16-bit code and
161 \IR{elf shared libraries} ELF, shared libraries
164 \IR{executable and linkable format} Executable and Linkable Format
165 \IR{extern, obj extensions to} \c{EXTERN}, \c{obj} extensions to
166 \IR{extern, rdf extensions to} \c{EXTERN}, \c{rdf} extensions to
167 \IR{floating-point, constants} floating-point, constants
168 \IR{floating-point, packed bcd constants} floating-point, packed BCD constants
170 \IR{freelink} FreeLink
171 \IR{functions, c calling convention} functions, C calling convention
172 \IR{functions, pascal calling convention} functions, Pascal calling
174 \IR{global, aoutb extensions to} \c{GLOBAL}, \c{aoutb} extensions to
175 \IR{global, elf extensions to} \c{GLOBAL}, \c{elf} extensions to
176 \IR{global, rdf extensions to} \c{GLOBAL}, \c{rdf} extensions to
178 \IR{got relocations} \c{GOT} relocations
179 \IR{gotoff relocation} \c{GOTOFF} relocations
180 \IR{gotpc relocation} \c{GOTPC} relocations
181 \IR{intel number formats} Intel number formats
182 \IR{linux, elf} Linux, ELF
183 \IR{linux, a.out} Linux, \c{a.out}
184 \IR{linux, as86} Linux, \c{as86}
185 \IR{logical and} logical AND
186 \IR{logical or} logical OR
187 \IR{logical xor} logical XOR
188 \IR{mach object file format} Mach, object file format
190 \IR{macho32} \c{macho32}
191 \IR{macho64} \c{macho64}
194 \IA{memory reference}{memory references}
196 \IA{misc directory}{misc subdirectory}
197 \IR{misc subdirectory} \c{misc} subdirectory
198 \IR{microsoft omf} Microsoft OMF
199 \IR{mmx registers} MMX registers
200 \IA{modr/m}{modr/m byte}
201 \IR{modr/m byte} ModR/M byte
203 \IR{ms-dos device drivers} MS-DOS device drivers
204 \IR{multipush} \c{multipush} macro
206 \IR{nasm version} NASM version
210 \IR{operating system} operating system
212 \IR{pascal calling convention}Pascal calling convention
213 \IR{passes} passes, assembly
218 \IR{plt} \c{PLT} relocations
219 \IA{pre-defining macros}{pre-define}
220 \IA{preprocessor expressions}{preprocessor, expressions}
221 \IA{preprocessor loops}{preprocessor, loops}
222 \IA{preprocessor variables}{preprocessor, variables}
223 \IA{rdoff subdirectory}{rdoff}
224 \IR{rdoff} \c{rdoff} subdirectory
225 \IR{relocatable dynamic object file format} Relocatable Dynamic
227 \IR{relocations, pic-specific} relocations, PIC-specific
228 \IA{repeating}{repeating code}
229 \IR{section alignment, in elf} section alignment, in \c{elf}
230 \IR{section alignment, in bin} section alignment, in \c{bin}
231 \IR{section alignment, in obj} section alignment, in \c{obj}
232 \IR{section alignment, in win32} section alignment, in \c{win32}
233 \IR{section, elf extensions to} \c{SECTION}, \c{elf} extensions to
234 \IR{section, win32 extensions to} \c{SECTION}, \c{win32} extensions to
235 \IR{segment alignment, in bin} segment alignment, in \c{bin}
236 \IR{segment alignment, in obj} segment alignment, in \c{obj}
237 \IR{segment, obj extensions to} \c{SEGMENT}, \c{elf} extensions to
238 \IR{segment names, borland pascal} segment names, Borland Pascal
239 \IR{shift command} \c{shift} command
241 \IR{sib byte} SIB byte
242 \IR{align, smart} \c{ALIGN}, smart
243 \IA{sectalign}{sectalign}
244 \IR{solaris x86} Solaris x86
245 \IA{standard section names}{standardized section names}
246 \IR{symbols, exporting from dlls} symbols, exporting from DLLs
247 \IR{symbols, importing from dlls} symbols, importing from DLLs
248 \IR{test subdirectory} \c{test} subdirectory
250 \IR{underscore, in c symbols} underscore, in C symbols
256 \IA{sco unix}{unix, sco}
257 \IR{unix, sco} Unix, SCO
258 \IA{unix source archive}{unix, source archive}
259 \IR{unix, source archive} Unix, source archive
260 \IA{unix system v}{unix, system v}
261 \IR{unix, system v} Unix, System V
262 \IR{unixware} UnixWare
264 \IR{version number of nasm} version number of NASM
265 \IR{visual c++} Visual C++
266 \IR{www page} WWW page
270 \IR{windows 95} Windows 95
271 \IR{windows nt} Windows NT
272 \# \IC{program entry point}{entry point, program}
273 \# \IC{program entry point}{start point, program}
274 \# \IC{MS-DOS device drivers}{device drivers, MS-DOS}
275 \# \IC{16-bit mode, versus 32-bit mode}{32-bit mode, versus 16-bit mode}
276 \# \IC{c symbol names}{symbol names, in C}
279 \C{intro} Introduction
281 \H{whatsnasm} What Is NASM?
283 The Netwide Assembler, NASM, is an 80x86 and x86-64 assembler designed
284 for portability and modularity. It supports a range of object file
285 formats, including Linux and \c{*BSD} \c{a.out}, \c{ELF}, \c{COFF},
286 \c{Mach-O}, Microsoft 16-bit \c{OBJ}, \c{Win32} and \c{Win64}. It will
287 also output plain binary files. Its syntax is designed to be simple
288 and easy to understand, similar to Intel's but less complex. It
289 supports all currently known x86 architectural extensions, and has
290 strong support for macros.
293 \S{yaasm} Why Yet Another Assembler?
295 The Netwide Assembler grew out of an idea on \i\c{comp.lang.asm.x86}
296 (or possibly \i\c{alt.lang.asm} - I forget which), which was
297 essentially that there didn't seem to be a good \e{free} x86-series
298 assembler around, and that maybe someone ought to write one.
300 \b \i\c{a86} is good, but not free, and in particular you don't get any
301 32-bit capability until you pay. It's DOS only, too.
303 \b \i\c{gas} is free, and ports over to DOS and Unix, but it's not
304 very good, since it's designed to be a back end to \i\c{gcc}, which
305 always feeds it correct code. So its error checking is minimal. Also,
306 its syntax is horrible, from the point of view of anyone trying to
307 actually \e{write} anything in it. Plus you can't write 16-bit code in
310 \b \i\c{as86} is specific to Minix and Linux, and (my version at least)
311 doesn't seem to have much (or any) documentation.
313 \b \i\c{MASM} isn't very good, and it's (was) expensive, and it runs only under
316 \b \i\c{TASM} is better, but still strives for MASM compatibility,
317 which means millions of directives and tons of red tape. And its syntax
318 is essentially MASM's, with the contradictions and quirks that
319 entails (although it sorts out some of those by means of Ideal mode.)
320 It's expensive too. And it's DOS-only.
322 So here, for your coding pleasure, is NASM. At present it's
323 still in prototype stage - we don't promise that it can outperform
324 any of these assemblers. But please, \e{please} send us bug reports,
325 fixes, helpful information, and anything else you can get your hands
326 on (and thanks to the many people who've done this already! You all
327 know who you are), and we'll improve it out of all recognition.
331 \S{legal} \i{License} Conditions
333 Please see the file \c{LICENSE}, supplied as part of any NASM
334 distribution archive, for the license conditions under which you may
335 use NASM. NASM is now under the so-called 2-clause BSD license, also
336 known as the simplified BSD license.
338 Copyright 1996-2011 the NASM Authors - All rights reserved.
340 Redistribution and use in source and binary forms, with or without
341 modification, are permitted provided that the following conditions are
344 \b Redistributions of source code must retain the above copyright
345 notice, this list of conditions and the following disclaimer.
347 \b Redistributions in binary form must reproduce the above copyright
348 notice, this list of conditions and the following disclaimer in the
349 documentation and/or other materials provided with the distribution.
351 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
352 CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
353 INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
354 MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
355 DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
356 CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
357 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
358 NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
359 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
360 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
361 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
362 OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
363 EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
366 \H{contact} Contact Information
368 The current version of NASM (since about 0.98.08) is maintained by a
369 team of developers, accessible through the \c{nasm-devel} mailing list
370 (see below for the link).
371 If you want to report a bug, please read \k{bugs} first.
373 NASM has a \i{website} at
374 \W{http://www.nasm.us/}\c{http://www.nasm.us/}. If it's not there,
377 \i{New releases}, \i{release candidates}, and \I{snapshots, daily
378 development}\i{daily development snapshots} of NASM are available from
379 the official web site.
381 Announcements are posted to
382 \W{news:comp.lang.asm.x86}\i\c{comp.lang.asm.x86},
384 \W{http://www.freshmeat.net/}\c{http://www.freshmeat.net/}.
386 If you want information about the current development status, please
387 subscribe to the \i\c{nasm-devel} email list; see link from the
391 \H{install} Installation
393 \S{instdos} \i{Installing} NASM under MS-\i{DOS} or Windows
395 Once you've obtained the appropriate archive for NASM,
396 \i\c{nasm-XXX-dos.zip} or \i\c{nasm-XXX-win32.zip} (where \c{XXX}
397 denotes the version number of NASM contained in the archive), unpack
398 it into its own directory (for example \c{c:\\nasm}).
400 The archive will contain a set of executable files: the NASM
401 executable file \i\c{nasm.exe}, the NDISASM executable file
402 \i\c{ndisasm.exe}, and possibly additional utilities to handle the
405 The only file NASM needs to run is its own executable, so copy
406 \c{nasm.exe} to a directory on your PATH, or alternatively edit
407 \i\c{autoexec.bat} to add the \c{nasm} directory to your
408 \i\c{PATH} (to do that under Windows XP, go to Start > Control Panel >
409 System > Advanced > Environment Variables; these instructions may work
410 under other versions of Windows as well.)
412 That's it - NASM is installed. You don't need the nasm directory
413 to be present to run NASM (unless you've added it to your \c{PATH}),
414 so you can delete it if you need to save space; however, you may
415 want to keep the documentation or test programs.
417 If you've downloaded the \i{DOS source archive}, \i\c{nasm-XXX.zip},
418 the \c{nasm} directory will also contain the full NASM \i{source
419 code}, and a selection of \i{Makefiles} you can (hopefully) use to
420 rebuild your copy of NASM from scratch. See the file \c{INSTALL} in
423 Note that a number of files are generated from other files by Perl
424 scripts. Although the NASM source distribution includes these
425 generated files, you will need to rebuild them (and hence, will need a
426 Perl interpreter) if you change insns.dat, standard.mac or the
427 documentation. It is possible future source distributions may not
428 include these files at all. Ports of \i{Perl} for a variety of
429 platforms, including DOS and Windows, are available from
430 \W{http://www.cpan.org/ports/}\i{www.cpan.org}.
433 \S{instdos} Installing NASM under \i{Unix}
435 Once you've obtained the \i{Unix source archive} for NASM,
436 \i\c{nasm-XXX.tar.gz} (where \c{XXX} denotes the version number of
437 NASM contained in the archive), unpack it into a directory such
438 as \c{/usr/local/src}. The archive, when unpacked, will create its
439 own subdirectory \c{nasm-XXX}.
441 NASM is an \I{Autoconf}\I\c{configure}auto-configuring package: once
442 you've unpacked it, \c{cd} to the directory it's been unpacked into
443 and type \c{./configure}. This shell script will find the best C
444 compiler to use for building NASM and set up \i{Makefiles}
447 Once NASM has auto-configured, you can type \i\c{make} to build the
448 \c{nasm} and \c{ndisasm} binaries, and then \c{make install} to
449 install them in \c{/usr/local/bin} and install the \i{man pages}
450 \i\c{nasm.1} and \i\c{ndisasm.1} in \c{/usr/local/man/man1}.
451 Alternatively, you can give options such as \c{--prefix} to the
452 configure script (see the file \i\c{INSTALL} for more details), or
453 install the programs yourself.
455 NASM also comes with a set of utilities for handling the \c{RDOFF}
456 custom object-file format, which are in the \i\c{rdoff} subdirectory
457 of the NASM archive. You can build these with \c{make rdf} and
458 install them with \c{make rdf_install}, if you want them.
461 \C{running} Running NASM
463 \H{syntax} NASM \i{Command-Line} Syntax
465 To assemble a file, you issue a command of the form
467 \c nasm -f <format> <filename> [-o <output>]
471 \c nasm -f elf myfile.asm
473 will assemble \c{myfile.asm} into an \c{ELF} object file \c{myfile.o}. And
475 \c nasm -f bin myfile.asm -o myfile.com
477 will assemble \c{myfile.asm} into a raw binary file \c{myfile.com}.
479 To produce a listing file, with the hex codes output from NASM
480 displayed on the left of the original sources, use the \c{-l} option
481 to give a listing file name, for example:
483 \c nasm -f coff myfile.asm -l myfile.lst
485 To get further usage instructions from NASM, try typing
489 As \c{-hf}, this will also list the available output file formats, and what they
492 If you use Linux but aren't sure whether your system is \c{a.out}
497 (in the directory in which you put the NASM binary when you
498 installed it). If it says something like
500 \c nasm: ELF 32-bit LSB executable i386 (386 and up) Version 1
502 then your system is \c{ELF}, and you should use the option \c{-f elf}
503 when you want NASM to produce Linux object files. If it says
505 \c nasm: Linux/i386 demand-paged executable (QMAGIC)
507 or something similar, your system is \c{a.out}, and you should use
508 \c{-f aout} instead (Linux \c{a.out} systems have long been obsolete,
509 and are rare these days.)
511 Like Unix compilers and assemblers, NASM is silent unless it
512 goes wrong: you won't see any output at all, unless it gives error
516 \S{opt-o} The \i\c{-o} Option: Specifying the Output File Name
518 NASM will normally choose the name of your output file for you;
519 precisely how it does this is dependent on the object file format.
520 For Microsoft object file formats (\c{obj}, \c{win32} and \c{win64}),
521 it will remove the \c{.asm} \i{extension} (or whatever extension you
522 like to use - NASM doesn't care) from your source file name and
523 substitute \c{.obj}. For Unix object file formats (\c{aout}, \c{as86},
524 \c{coff}, \c{elf32}, \c{elf64}, \c{ieee}, \c{macho32} and \c{macho64})
525 it will substitute \c{.o}. For \c{dbg}, \c{rdf}, \c{ith} and \c{srec},
526 it will use \c{.dbg}, \c{.rdf}, \c{.ith} and \c{.srec}, respectively,
527 and for the \c{bin} format it will simply remove the extension, so
528 that \c{myfile.asm} produces the output file \c{myfile}.
530 If the output file already exists, NASM will overwrite it, unless it
531 has the same name as the input file, in which case it will give a
532 warning and use \i\c{nasm.out} as the output file name instead.
534 For situations in which this behaviour is unacceptable, NASM
535 provides the \c{-o} command-line option, which allows you to specify
536 your desired output file name. You invoke \c{-o} by following it
537 with the name you wish for the output file, either with or without
538 an intervening space. For example:
540 \c nasm -f bin program.asm -o program.com
541 \c nasm -f bin driver.asm -odriver.sys
543 Note that this is a small o, and is different from a capital O , which
544 is used to specify the number of optimisation passes required. See \k{opt-O}.
547 \S{opt-f} The \i\c{-f} Option: Specifying the \i{Output File Format}
549 If you do not supply the \c{-f} option to NASM, it will choose an
550 output file format for you itself. In the distribution versions of
551 NASM, the default is always \i\c{bin}; if you've compiled your own
552 copy of NASM, you can redefine \i\c{OF_DEFAULT} at compile time and
553 choose what you want the default to be.
555 Like \c{-o}, the intervening space between \c{-f} and the output
556 file format is optional; so \c{-f elf} and \c{-felf} are both valid.
558 A complete list of the available output file formats can be given by
559 issuing the command \i\c{nasm -hf}.
562 \S{opt-l} The \i\c{-l} Option: Generating a \i{Listing File}
564 If you supply the \c{-l} option to NASM, followed (with the usual
565 optional space) by a file name, NASM will generate a
566 \i{source-listing file} for you, in which addresses and generated
567 code are listed on the left, and the actual source code, with
568 expansions of multi-line macros (except those which specifically
569 request no expansion in source listings: see \k{nolist}) on the
572 \c nasm -f elf myfile.asm -l myfile.lst
574 If a list file is selected, you may turn off listing for a
575 section of your source with \c{[list -]}, and turn it back on
576 with \c{[list +]}, (the default, obviously). There is no "user
577 form" (without the brackets). This can be used to list only
578 sections of interest, avoiding excessively long listings.
581 \S{opt-M} The \i\c{-M} Option: Generate \i{Makefile Dependencies}
583 This option can be used to generate makefile dependencies on stdout.
584 This can be redirected to a file for further processing. For example:
586 \c nasm -M myfile.asm > myfile.dep
589 \S{opt-MG} The \i\c{-MG} Option: Generate \i{Makefile Dependencies}
591 This option can be used to generate makefile dependencies on stdout.
592 This differs from the \c{-M} option in that if a nonexisting file is
593 encountered, it is assumed to be a generated file and is added to the
594 dependency list without a prefix.
597 \S{opt-MF} The \i\c\{-MF} Option: Set Makefile Dependency File
599 This option can be used with the \c{-M} or \c{-MG} options to send the
600 output to a file, rather than to stdout. For example:
602 \c nasm -M -MF myfile.dep myfile.asm
605 \S{opt-MD} The \i\c{-MD} Option: Assemble and Generate Dependencies
607 The \c{-MD} option acts as the combination of the \c{-M} and \c{-MF}
608 options (i.e. a filename has to be specified.) However, unlike the
609 \c{-M} or \c{-MG} options, \c{-MD} does \e{not} inhibit the normal
610 operation of the assembler. Use this to automatically generate
611 updated dependencies with every assembly session. For example:
613 \c nasm -f elf -o myfile.o -MD myfile.dep myfile.asm
616 \S{opt-MT} The \i\c{-MT} Option: Dependency Target Name
618 The \c{-MT} option can be used to override the default name of the
619 dependency target. This is normally the same as the output filename,
620 specified by the \c{-o} option.
623 \S{opt-MQ} The \i\c{-MQ} Option: Dependency Target Name (Quoted)
625 The \c{-MQ} option acts as the \c{-MT} option, except it tries to
626 quote characters that have special meaning in Makefile syntax. This
627 is not foolproof, as not all characters with special meaning are
631 \S{opt-MP} The \i\c{-MP} Option: Emit phony targets
633 When used with any of the dependency generation options, the \c{-MP}
634 option causes NASM to emit a phony target without dependencies for
635 each header file. This prevents Make from complaining if a header
636 file has been removed.
639 \S{opt-F} The \i\c{-F} Option: Selecting a \i{Debug Information Format}
641 This option is used to select the format of the debug information
642 emitted into the output file, to be used by a debugger (or \e{will}
643 be). Prior to version 2.03.01, the use of this switch did \e{not} enable
644 output of the selected debug info format. Use \c{-g}, see \k{opt-g},
645 to enable output. Versions 2.03.01 and later automatically enable \c{-g}
646 if \c{-F} is specified.
648 A complete list of the available debug file formats for an output
649 format can be seen by issuing the command \c{nasm -f <format> -y}. Not
650 all output formats currently support debugging output. See \k{opt-y}.
652 This should not be confused with the \c{-f dbg} output format option which
653 is not built into NASM by default. For information on how
654 to enable it when building from the sources, see \k{dbgfmt}.
657 \S{opt-g} The \i\c{-g} Option: Enabling \i{Debug Information}.
659 This option can be used to generate debugging information in the specified
660 format. See \k{opt-F}. Using \c{-g} without \c{-F} results in emitting
661 debug info in the default format, if any, for the selected output format.
662 If no debug information is currently implemented in the selected output
663 format, \c{-g} is \e{silently ignored}.
666 \S{opt-X} The \i\c{-X} Option: Selecting an \i{Error Reporting Format}
668 This option can be used to select an error reporting format for any
669 error messages that might be produced by NASM.
671 Currently, two error reporting formats may be selected. They are
672 the \c{-Xvc} option and the \c{-Xgnu} option. The GNU format is
673 the default and looks like this:
675 \c filename.asm:65: error: specific error message
677 where \c{filename.asm} is the name of the source file in which the
678 error was detected, \c{65} is the source file line number on which
679 the error was detected, \c{error} is the severity of the error (this
680 could be \c{warning}), and \c{specific error message} is a more
681 detailed text message which should help pinpoint the exact problem.
683 The other format, specified by \c{-Xvc} is the style used by Microsoft
684 Visual C++ and some other programs. It looks like this:
686 \c filename.asm(65) : error: specific error message
688 where the only difference is that the line number is in parentheses
689 instead of being delimited by colons.
691 See also the \c{Visual C++} output format, \k{win32fmt}.
693 \S{opt-Z} The \i\c{-Z} Option: Send Errors to a File
695 Under \I{DOS}\c{MS-DOS} it can be difficult (though there are ways) to
696 redirect the standard-error output of a program to a file. Since
697 NASM usually produces its warning and \i{error messages} on
698 \i\c{stderr}, this can make it hard to capture the errors if (for
699 example) you want to load them into an editor.
701 NASM therefore provides the \c{-Z} option, taking a filename argument
702 which causes errors to be sent to the specified files rather than
703 standard error. Therefore you can \I{redirecting errors}redirect
704 the errors into a file by typing
706 \c nasm -Z myfile.err -f obj myfile.asm
708 In earlier versions of NASM, this option was called \c{-E}, but it was
709 changed since \c{-E} is an option conventionally used for
710 preprocessing only, with disastrous results. See \k{opt-E}.
712 \S{opt-s} The \i\c{-s} Option: Send Errors to \i\c{stdout}
714 The \c{-s} option redirects \i{error messages} to \c{stdout} rather
715 than \c{stderr}, so it can be redirected under \I{DOS}\c{MS-DOS}. To
716 assemble the file \c{myfile.asm} and pipe its output to the \c{more}
717 program, you can type:
719 \c nasm -s -f obj myfile.asm | more
721 See also the \c{-Z} option, \k{opt-Z}.
724 \S{opt-i} The \i\c{-i}\I\c{-I} Option: Include File Search Directories
726 When NASM sees the \i\c{%include} or \i\c{%pathsearch} directive in a
727 source file (see \k{include}, \k{pathsearch} or \k{incbin}), it will
728 search for the given file not only in the current directory, but also
729 in any directories specified on the command line by the use of the
730 \c{-i} option. Therefore you can include files from a \i{macro
731 library}, for example, by typing
733 \c nasm -ic:\macrolib\ -f obj myfile.asm
735 (As usual, a space between \c{-i} and the path name is allowed, and
738 NASM, in the interests of complete source-code portability, does not
739 understand the file naming conventions of the OS it is running on;
740 the string you provide as an argument to the \c{-i} option will be
741 prepended exactly as written to the name of the include file.
742 Therefore the trailing backslash in the above example is necessary.
743 Under Unix, a trailing forward slash is similarly necessary.
745 (You can use this to your advantage, if you're really \i{perverse},
746 by noting that the option \c{-ifoo} will cause \c{%include "bar.i"}
747 to search for the file \c{foobar.i}...)
749 If you want to define a \e{standard} \i{include search path},
750 similar to \c{/usr/include} on Unix systems, you should place one or
751 more \c{-i} directives in the \c{NASMENV} environment variable (see
754 For Makefile compatibility with many C compilers, this option can also
755 be specified as \c{-I}.
758 \S{opt-p} The \i\c{-p}\I\c{-P} Option: \I{pre-including files}Pre-Include a File
760 \I\c{%include}NASM allows you to specify files to be
761 \e{pre-included} into your source file, by the use of the \c{-p}
764 \c nasm myfile.asm -p myinc.inc
766 is equivalent to running \c{nasm myfile.asm} and placing the
767 directive \c{%include "myinc.inc"} at the start of the file.
769 For consistency with the \c{-I}, \c{-D} and \c{-U} options, this
770 option can also be specified as \c{-P}.
773 \S{opt-d} The \i\c{-d}\I\c{-D} Option: \I{pre-defining macros}Pre-Define a Macro
775 \I\c{%define}Just as the \c{-p} option gives an alternative to placing
776 \c{%include} directives at the start of a source file, the \c{-d}
777 option gives an alternative to placing a \c{%define} directive. You
780 \c nasm myfile.asm -dFOO=100
782 as an alternative to placing the directive
786 at the start of the file. You can miss off the macro value, as well:
787 the option \c{-dFOO} is equivalent to coding \c{%define FOO}. This
788 form of the directive may be useful for selecting \i{assembly-time
789 options} which are then tested using \c{%ifdef}, for example
792 For Makefile compatibility with many C compilers, this option can also
793 be specified as \c{-D}.
796 \S{opt-u} The \i\c{-u}\I\c{-U} Option: \I{Undefining macros}Undefine a Macro
798 \I\c{%undef}The \c{-u} option undefines a macro that would otherwise
799 have been pre-defined, either automatically or by a \c{-p} or \c{-d}
800 option specified earlier on the command lines.
802 For example, the following command line:
804 \c nasm myfile.asm -dFOO=100 -uFOO
806 would result in \c{FOO} \e{not} being a predefined macro in the
807 program. This is useful to override options specified at a different
810 For Makefile compatibility with many C compilers, this option can also
811 be specified as \c{-U}.
814 \S{opt-E} The \i\c{-E}\I{-e} Option: Preprocess Only
816 NASM allows the \i{preprocessor} to be run on its own, up to a
817 point. Using the \c{-E} option (which requires no arguments) will
818 cause NASM to preprocess its input file, expand all the macro
819 references, remove all the comments and preprocessor directives, and
820 print the resulting file on standard output (or save it to a file,
821 if the \c{-o} option is also used).
823 This option cannot be applied to programs which require the
824 preprocessor to evaluate \I{preprocessor expressions}\i{expressions}
825 which depend on the values of symbols: so code such as
827 \c %assign tablesize ($-tablestart)
829 will cause an error in \i{preprocess-only mode}.
831 For compatiblity with older version of NASM, this option can also be
832 written \c{-e}. \c{-E} in older versions of NASM was the equivalent
833 of the current \c{-Z} option, \k{opt-Z}.
835 \S{opt-a} The \i\c{-a} Option: Don't Preprocess At All
837 If NASM is being used as the back end to a compiler, it might be
838 desirable to \I{suppressing preprocessing}suppress preprocessing
839 completely and assume the compiler has already done it, to save time
840 and increase compilation speeds. The \c{-a} option, requiring no
841 argument, instructs NASM to replace its powerful \i{preprocessor}
842 with a \i{stub preprocessor} which does nothing.
845 \S{opt-O} The \i\c{-O} Option: Specifying \i{Multipass Optimization}
847 Using the \c{-O} option, you can tell NASM to carry out different
848 levels of optimization. The syntax is:
850 \b \c{-O0}: No optimization. All operands take their long forms,
851 if a short form is not specified, except conditional jumps.
852 This is intended to match NASM 0.98 behavior.
854 \b \c{-O1}: Minimal optimization. As above, but immediate operands
855 which will fit in a signed byte are optimized,
856 unless the long form is specified. Conditional jumps default
857 to the long form unless otherwise specified.
859 \b \c{-Ox} (where \c{x} is the actual letter \c{x}): Multipass optimization.
860 Minimize branch offsets and signed immediate bytes,
861 overriding size specification unless the \c{strict} keyword
862 has been used (see \k{strict}). For compatibility with earlier
863 releases, the letter \c{x} may also be any number greater than
864 one. This number has no effect on the actual number of passes.
866 The \c{-Ox} mode is recommended for most uses, and is the default
869 Note that this is a capital \c{O}, and is different from a small \c{o}, which
870 is used to specify the output file name. See \k{opt-o}.
873 \S{opt-t} The \i\c{-t} Option: Enable TASM Compatibility Mode
875 NASM includes a limited form of compatibility with Borland's \i\c{TASM}.
876 When NASM's \c{-t} option is used, the following changes are made:
878 \b local labels may be prefixed with \c{@@} instead of \c{.}
880 \b size override is supported within brackets. In TASM compatible mode,
881 a size override inside square brackets changes the size of the operand,
882 and not the address type of the operand as it does in NASM syntax. E.g.
883 \c{mov eax,[DWORD val]} is valid syntax in TASM compatibility mode.
884 Note that you lose the ability to override the default address type for
887 \b unprefixed forms of some directives supported (\c{arg}, \c{elif},
888 \c{else}, \c{endif}, \c{if}, \c{ifdef}, \c{ifdifi}, \c{ifndef},
889 \c{include}, \c{local})
891 \S{opt-w} The \i\c{-w} and \i\c{-W} Options: Enable or Disable Assembly \i{Warnings}
893 NASM can observe many conditions during the course of assembly which
894 are worth mentioning to the user, but not a sufficiently severe
895 error to justify NASM refusing to generate an output file. These
896 conditions are reported like errors, but come up with the word
897 `warning' before the message. Warnings do not prevent NASM from
898 generating an output file and returning a success status to the
901 Some conditions are even less severe than that: they are only
902 sometimes worth mentioning to the user. Therefore NASM supports the
903 \c{-w} command-line option, which enables or disables certain
904 classes of assembly warning. Such warning classes are described by a
905 name, for example \c{orphan-labels}; you can enable warnings of
906 this class by the command-line option \c{-w+orphan-labels} and
907 disable it by \c{-w-orphan-labels}.
909 The \i{suppressible warning} classes are:
911 \b \i\c{macro-params} covers warnings about \i{multi-line macros}
912 being invoked with the wrong number of parameters. This warning
913 class is enabled by default; see \k{mlmacover} for an example of why
914 you might want to disable it.
916 \b \i\c{macro-selfref} warns if a macro references itself. This
917 warning class is disabled by default.
919 \b\i\c{macro-defaults} warns when a macro has more default
920 parameters than optional parameters. This warning class
921 is enabled by default; see \k{mlmacdef} for why you might want to disable it.
923 \b \i\c{orphan-labels} covers warnings about source lines which
924 contain no instruction but define a label without a trailing colon.
925 NASM warns about this somewhat obscure condition by default;
926 see \k{syntax} for more information.
928 \b \i\c{number-overflow} covers warnings about numeric constants which
929 don't fit in 64 bits. This warning class is enabled by default.
931 \b \i\c{gnu-elf-extensions} warns if 8-bit or 16-bit relocations
932 are used in \c{-f elf} format. The GNU extensions allow this.
933 This warning class is disabled by default.
935 \b \i\c{float-overflow} warns about floating point overflow.
938 \b \i\c{float-denorm} warns about floating point denormals.
941 \b \i\c{float-underflow} warns about floating point underflow.
944 \b \i\c{float-toolong} warns about too many digits in floating-point numbers.
947 \b \i\c{user} controls \c{%warning} directives (see \k{pperror}).
950 \b \i\c{error} causes warnings to be treated as errors. Disabled by
953 \b \i\c{all} is an alias for \e{all} suppressible warning classes (not
954 including \c{error}). Thus, \c{-w+all} enables all available warnings.
956 In addition, you can set warning classes across sections.
957 Warning classes may be enabled with \i\c{[warning +warning-name]},
958 disabled with \i\c{[warning -warning-name]} or reset to their
959 original value with \i\c{[warning *warning-name]}. No "user form"
960 (without the brackets) exists.
962 Since version 2.00, NASM has also supported the gcc-like syntax
963 \c{-Wwarning} and \c{-Wno-warning} instead of \c{-w+warning} and
964 \c{-w-warning}, respectively.
967 \S{opt-v} The \i\c{-v} Option: Display \i{Version} Info
969 Typing \c{NASM -v} will display the version of NASM which you are using,
970 and the date on which it was compiled.
972 You will need the version number if you report a bug.
974 \S{opt-y} The \i\c{-y} Option: Display Available Debug Info Formats
976 Typing \c{nasm -f <option> -y} will display a list of the available
977 debug info formats for the given output format. The default format
978 is indicated by an asterisk. For example:
982 \c valid debug formats for 'elf32' output format are
983 \c ('*' denotes default):
984 \c * stabs ELF32 (i386) stabs debug format for Linux
985 \c dwarf elf32 (i386) dwarf debug format for Linux
988 \S{opt-pfix} The \i\c{--prefix} and \i\c{--postfix} Options.
990 The \c{--prefix} and \c{--postfix} options prepend or append
991 (respectively) the given argument to all \c{global} or
992 \c{extern} variables. E.g. \c{--prefix _} will prepend the
993 underscore to all global and external variables, as C sometimes
994 (but not always) likes it.
997 \S{nasmenv} The \i\c{NASMENV} \i{Environment} Variable
999 If you define an environment variable called \c{NASMENV}, the program
1000 will interpret it as a list of extra command-line options, which are
1001 processed before the real command line. You can use this to define
1002 standard search directories for include files, by putting \c{-i}
1003 options in the \c{NASMENV} variable.
1005 The value of the variable is split up at white space, so that the
1006 value \c{-s -ic:\\nasmlib\\} will be treated as two separate options.
1007 However, that means that the value \c{-dNAME="my name"} won't do
1008 what you might want, because it will be split at the space and the
1009 NASM command-line processing will get confused by the two
1010 nonsensical words \c{-dNAME="my} and \c{name"}.
1012 To get round this, NASM provides a feature whereby, if you begin the
1013 \c{NASMENV} environment variable with some character that isn't a minus
1014 sign, then NASM will treat this character as the \i{separator
1015 character} for options. So setting the \c{NASMENV} variable to the
1016 value \c{!-s!-ic:\\nasmlib\\} is equivalent to setting it to \c{-s
1017 -ic:\\nasmlib\\}, but \c{!-dNAME="my name"} will work.
1019 This environment variable was previously called \c{NASM}. This was
1020 changed with version 0.98.31.
1023 \H{qstart} \i{Quick Start} for \i{MASM} Users
1025 If you're used to writing programs with MASM, or with \i{TASM} in
1026 MASM-compatible (non-Ideal) mode, or with \i\c{a86}, this section
1027 attempts to outline the major differences between MASM's syntax and
1028 NASM's. If you're not already used to MASM, it's probably worth
1029 skipping this section.
1032 \S{qscs} NASM Is \I{case sensitivity}Case-Sensitive
1034 One simple difference is that NASM is case-sensitive. It makes a
1035 difference whether you call your label \c{foo}, \c{Foo} or \c{FOO}.
1036 If you're assembling to \c{DOS} or \c{OS/2} \c{.OBJ} files, you can
1037 invoke the \i\c{UPPERCASE} directive (documented in \k{objfmt}) to
1038 ensure that all symbols exported to other code modules are forced
1039 to be upper case; but even then, \e{within} a single module, NASM
1040 will distinguish between labels differing only in case.
1043 \S{qsbrackets} NASM Requires \i{Square Brackets} For \i{Memory References}
1045 NASM was designed with simplicity of syntax in mind. One of the
1046 \i{design goals} of NASM is that it should be possible, as far as is
1047 practical, for the user to look at a single line of NASM code
1048 and tell what opcode is generated by it. You can't do this in MASM:
1049 if you declare, for example,
1054 then the two lines of code
1059 generate completely different opcodes, despite having
1060 identical-looking syntaxes.
1062 NASM avoids this undesirable situation by having a much simpler
1063 syntax for memory references. The rule is simply that any access to
1064 the \e{contents} of a memory location requires square brackets
1065 around the address, and any access to the \e{address} of a variable
1066 doesn't. So an instruction of the form \c{mov ax,foo} will
1067 \e{always} refer to a compile-time constant, whether it's an \c{EQU}
1068 or the address of a variable; and to access the \e{contents} of the
1069 variable \c{bar}, you must code \c{mov ax,[bar]}.
1071 This also means that NASM has no need for MASM's \i\c{OFFSET}
1072 keyword, since the MASM code \c{mov ax,offset bar} means exactly the
1073 same thing as NASM's \c{mov ax,bar}. If you're trying to get
1074 large amounts of MASM code to assemble sensibly under NASM, you
1075 can always code \c{%idefine offset} to make the preprocessor treat
1076 the \c{OFFSET} keyword as a no-op.
1078 This issue is even more confusing in \i\c{a86}, where declaring a
1079 label with a trailing colon defines it to be a `label' as opposed to
1080 a `variable' and causes \c{a86} to adopt NASM-style semantics; so in
1081 \c{a86}, \c{mov ax,var} has different behaviour depending on whether
1082 \c{var} was declared as \c{var: dw 0} (a label) or \c{var dw 0} (a
1083 word-size variable). NASM is very simple by comparison:
1084 \e{everything} is a label.
1086 NASM, in the interests of simplicity, also does not support the
1087 \i{hybrid syntaxes} supported by MASM and its clones, such as
1088 \c{mov ax,table[bx]}, where a memory reference is denoted by one
1089 portion outside square brackets and another portion inside. The
1090 correct syntax for the above is \c{mov ax,[table+bx]}. Likewise,
1091 \c{mov ax,es:[di]} is wrong and \c{mov ax,[es:di]} is right.
1094 \S{qstypes} NASM Doesn't Store \i{Variable Types}
1096 NASM, by design, chooses not to remember the types of variables you
1097 declare. Whereas MASM will remember, on seeing \c{var dw 0}, that
1098 you declared \c{var} as a word-size variable, and will then be able
1099 to fill in the \i{ambiguity} in the size of the instruction \c{mov
1100 var,2}, NASM will deliberately remember nothing about the symbol
1101 \c{var} except where it begins, and so you must explicitly code
1102 \c{mov word [var],2}.
1104 For this reason, NASM doesn't support the \c{LODS}, \c{MOVS},
1105 \c{STOS}, \c{SCAS}, \c{CMPS}, \c{INS}, or \c{OUTS} instructions,
1106 but only supports the forms such as \c{LODSB}, \c{MOVSW}, and
1107 \c{SCASD}, which explicitly specify the size of the components of
1108 the strings being manipulated.
1111 \S{qsassume} NASM Doesn't \i\c{ASSUME}
1113 As part of NASM's drive for simplicity, it also does not support the
1114 \c{ASSUME} directive. NASM will not keep track of what values you
1115 choose to put in your segment registers, and will never
1116 \e{automatically} generate a \i{segment override} prefix.
1119 \S{qsmodel} NASM Doesn't Support \i{Memory Models}
1121 NASM also does not have any directives to support different 16-bit
1122 memory models. The programmer has to keep track of which functions
1123 are supposed to be called with a \i{far call} and which with a
1124 \i{near call}, and is responsible for putting the correct form of
1125 \c{RET} instruction (\c{RETN} or \c{RETF}; NASM accepts \c{RET}
1126 itself as an alternate form for \c{RETN}); in addition, the
1127 programmer is responsible for coding CALL FAR instructions where
1128 necessary when calling \e{external} functions, and must also keep
1129 track of which external variable definitions are far and which are
1133 \S{qsfpu} \i{Floating-Point} Differences
1135 NASM uses different names to refer to floating-point registers from
1136 MASM: where MASM would call them \c{ST(0)}, \c{ST(1)} and so on, and
1137 \i\c{a86} would call them simply \c{0}, \c{1} and so on, NASM
1138 chooses to call them \c{st0}, \c{st1} etc.
1140 As of version 0.96, NASM now treats the instructions with
1141 \i{`nowait'} forms in the same way as MASM-compatible assemblers.
1142 The idiosyncratic treatment employed by 0.95 and earlier was based
1143 on a misunderstanding by the authors.
1146 \S{qsother} Other Differences
1148 For historical reasons, NASM uses the keyword \i\c{TWORD} where MASM
1149 and compatible assemblers use \i\c{TBYTE}.
1151 NASM does not declare \i{uninitialized storage} in the same way as
1152 MASM: where a MASM programmer might use \c{stack db 64 dup (?)},
1153 NASM requires \c{stack resb 64}, intended to be read as `reserve 64
1154 bytes'. For a limited amount of compatibility, since NASM treats
1155 \c{?} as a valid character in symbol names, you can code \c{? equ 0}
1156 and then writing \c{dw ?} will at least do something vaguely useful.
1157 \I\c{RESB}\i\c{DUP} is still not a supported syntax, however.
1159 In addition to all of this, macros and directives work completely
1160 differently to MASM. See \k{preproc} and \k{directive} for further
1164 \C{lang} The NASM Language
1166 \H{syntax} Layout of a NASM Source Line
1168 Like most assemblers, each NASM source line contains (unless it
1169 is a macro, a preprocessor directive or an assembler directive: see
1170 \k{preproc} and \k{directive}) some combination of the four fields
1172 \c label: instruction operands ; comment
1174 As usual, most of these fields are optional; the presence or absence
1175 of any combination of a label, an instruction and a comment is allowed.
1176 Of course, the operand field is either required or forbidden by the
1177 presence and nature of the instruction field.
1179 NASM uses backslash (\\) as the line continuation character; if a line
1180 ends with backslash, the next line is considered to be a part of the
1181 backslash-ended line.
1183 NASM places no restrictions on white space within a line: labels may
1184 have white space before them, or instructions may have no space
1185 before them, or anything. The \i{colon} after a label is also
1186 optional. (Note that this means that if you intend to code \c{lodsb}
1187 alone on a line, and type \c{lodab} by accident, then that's still a
1188 valid source line which does nothing but define a label. Running
1189 NASM with the command-line option
1190 \I{orphan-labels}\c{-w+orphan-labels} will cause it to warn you if
1191 you define a label alone on a line without a \i{trailing colon}.)
1193 \i{Valid characters} in labels are letters, numbers, \c{_}, \c{$},
1194 \c{#}, \c{@}, \c{~}, \c{.}, and \c{?}. The only characters which may
1195 be used as the \e{first} character of an identifier are letters,
1196 \c{.} (with special meaning: see \k{locallab}), \c{_} and \c{?}.
1197 An identifier may also be prefixed with a \I{$, prefix}\c{$} to
1198 indicate that it is intended to be read as an identifier and not a
1199 reserved word; thus, if some other module you are linking with
1200 defines a symbol called \c{eax}, you can refer to \c{$eax} in NASM
1201 code to distinguish the symbol from the register. Maximum length of
1202 an identifier is 4095 characters.
1204 The instruction field may contain any machine instruction: Pentium
1205 and P6 instructions, FPU instructions, MMX instructions and even
1206 undocumented instructions are all supported. The instruction may be
1207 prefixed by \c{LOCK}, \c{REP}, \c{REPE}/\c{REPZ} or
1208 \c{REPNE}/\c{REPNZ}, in the usual way. Explicit \I{address-size
1209 prefixes}address-size and \i{operand-size prefixes} \i\c{A16},
1210 \i\c{A32}, \i\c{A64}, \i\c{O16} and \i\c{O32}, \i\c{O64} are provided - one example of their use
1211 is given in \k{mixsize}. You can also use the name of a \I{segment
1212 override}segment register as an instruction prefix: coding
1213 \c{es mov [bx],ax} is equivalent to coding \c{mov [es:bx],ax}. We
1214 recommend the latter syntax, since it is consistent with other
1215 syntactic features of the language, but for instructions such as
1216 \c{LODSB}, which has no operands and yet can require a segment
1217 override, there is no clean syntactic way to proceed apart from
1220 An instruction is not required to use a prefix: prefixes such as
1221 \c{CS}, \c{A32}, \c{LOCK} or \c{REPE} can appear on a line by
1222 themselves, and NASM will just generate the prefix bytes.
1224 In addition to actual machine instructions, NASM also supports a
1225 number of pseudo-instructions, described in \k{pseudop}.
1227 Instruction \i{operands} may take a number of forms: they can be
1228 registers, described simply by the register name (e.g. \c{ax},
1229 \c{bp}, \c{ebx}, \c{cr0}: NASM does not use the \c{gas}-style
1230 syntax in which register names must be prefixed by a \c{%} sign), or
1231 they can be \i{effective addresses} (see \k{effaddr}), constants
1232 (\k{const}) or expressions (\k{expr}).
1234 For x87 \i{floating-point} instructions, NASM accepts a wide range of
1235 syntaxes: you can use two-operand forms like MASM supports, or you
1236 can use NASM's native single-operand forms in most cases.
1238 \# all forms of each supported instruction are given in
1240 For example, you can code:
1242 \c fadd st1 ; this sets st0 := st0 + st1
1243 \c fadd st0,st1 ; so does this
1245 \c fadd st1,st0 ; this sets st1 := st1 + st0
1246 \c fadd to st1 ; so does this
1248 Almost any x87 floating-point instruction that references memory must
1249 use one of the prefixes \i\c{DWORD}, \i\c{QWORD} or \i\c{TWORD} to
1250 indicate what size of \i{memory operand} it refers to.
1253 \H{pseudop} \i{Pseudo-Instructions}
1255 Pseudo-instructions are things which, though not real x86 machine
1256 instructions, are used in the instruction field anyway because that's
1257 the most convenient place to put them. The current pseudo-instructions
1258 are \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, \i\c{DO} and
1259 \i\c{DY}; their \i{uninitialized} counterparts \i\c{RESB}, \i\c{RESW},
1260 \i\c{RESD}, \i\c{RESQ}, \i\c{REST}, \i\c{RESO} and \i\c{RESY}; the
1261 \i\c{INCBIN} command, the \i\c{EQU} command, and the \i\c{TIMES}
1265 \S{db} \c{DB} and Friends: Declaring Initialized Data
1267 \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, \i\c{DO} and
1268 \i\c{DY} are used, much as in MASM, to declare initialized data in the
1269 output file. They can be invoked in a wide range of ways:
1270 \I{floating-point}\I{character constant}\I{string constant}
1272 \c db 0x55 ; just the byte 0x55
1273 \c db 0x55,0x56,0x57 ; three bytes in succession
1274 \c db 'a',0x55 ; character constants are OK
1275 \c db 'hello',13,10,'$' ; so are string constants
1276 \c dw 0x1234 ; 0x34 0x12
1277 \c dw 'a' ; 0x61 0x00 (it's just a number)
1278 \c dw 'ab' ; 0x61 0x62 (character constant)
1279 \c dw 'abc' ; 0x61 0x62 0x63 0x00 (string)
1280 \c dd 0x12345678 ; 0x78 0x56 0x34 0x12
1281 \c dd 1.234567e20 ; floating-point constant
1282 \c dq 0x123456789abcdef0 ; eight byte constant
1283 \c dq 1.234567e20 ; double-precision float
1284 \c dt 1.234567e20 ; extended-precision float
1286 \c{DT}, \c{DO} and \c{DY} do not accept \i{numeric constants} as operands.
1289 \S{resb} \c{RESB} and Friends: Declaring \i{Uninitialized} Data
1291 \i\c{RESB}, \i\c{RESW}, \i\c{RESD}, \i\c{RESQ}, \i\c{REST}, \i\c{RESO}
1292 and \i\c{RESY} are designed to be used in the BSS section of a module:
1293 they declare \e{uninitialized} storage space. Each takes a single
1294 operand, which is the number of bytes, words, doublewords or whatever
1295 to reserve. As stated in \k{qsother}, NASM does not support the
1296 MASM/TASM syntax of reserving uninitialized space by writing
1297 \I\c{?}\c{DW ?} or similar things: this is what it does instead. The
1298 operand to a \c{RESB}-type pseudo-instruction is a \i\e{critical
1299 expression}: see \k{crit}.
1303 \c buffer: resb 64 ; reserve 64 bytes
1304 \c wordvar: resw 1 ; reserve a word
1305 \c realarray resq 10 ; array of ten reals
1306 \c ymmval: resy 1 ; one YMM register
1308 \S{incbin} \i\c{INCBIN}: Including External \i{Binary Files}
1310 \c{INCBIN} is borrowed from the old Amiga assembler \i{DevPac}: it
1311 includes a binary file verbatim into the output file. This can be
1312 handy for (for example) including \i{graphics} and \i{sound} data
1313 directly into a game executable file. It can be called in one of
1316 \c incbin "file.dat" ; include the whole file
1317 \c incbin "file.dat",1024 ; skip the first 1024 bytes
1318 \c incbin "file.dat",1024,512 ; skip the first 1024, and
1319 \c ; actually include at most 512
1321 \c{INCBIN} is both a directive and a standard macro; the standard
1322 macro version searches for the file in the include file search path
1323 and adds the file to the dependency lists. This macro can be
1324 overridden if desired.
1327 \S{equ} \i\c{EQU}: Defining Constants
1329 \c{EQU} defines a symbol to a given constant value: when \c{EQU} is
1330 used, the source line must contain a label. The action of \c{EQU} is
1331 to define the given label name to the value of its (only) operand.
1332 This definition is absolute, and cannot change later. So, for
1335 \c message db 'hello, world'
1336 \c msglen equ $-message
1338 defines \c{msglen} to be the constant 12. \c{msglen} may not then be
1339 redefined later. This is not a \i{preprocessor} definition either:
1340 the value of \c{msglen} is evaluated \e{once}, using the value of
1341 \c{$} (see \k{expr} for an explanation of \c{$}) at the point of
1342 definition, rather than being evaluated wherever it is referenced
1343 and using the value of \c{$} at the point of reference.
1346 \S{times} \i\c{TIMES}: \i{Repeating} Instructions or Data
1348 The \c{TIMES} prefix causes the instruction to be assembled multiple
1349 times. This is partly present as NASM's equivalent of the \i\c{DUP}
1350 syntax supported by \i{MASM}-compatible assemblers, in that you can
1353 \c zerobuf: times 64 db 0
1355 or similar things; but \c{TIMES} is more versatile than that. The
1356 argument to \c{TIMES} is not just a numeric constant, but a numeric
1357 \e{expression}, so you can do things like
1359 \c buffer: db 'hello, world'
1360 \c times 64-$+buffer db ' '
1362 which will store exactly enough spaces to make the total length of
1363 \c{buffer} up to 64. Finally, \c{TIMES} can be applied to ordinary
1364 instructions, so you can code trivial \i{unrolled loops} in it:
1368 Note that there is no effective difference between \c{times 100 resb
1369 1} and \c{resb 100}, except that the latter will be assembled about
1370 100 times faster due to the internal structure of the assembler.
1372 The operand to \c{TIMES} is a critical expression (\k{crit}).
1374 Note also that \c{TIMES} can't be applied to \i{macros}: the reason
1375 for this is that \c{TIMES} is processed after the macro phase, which
1376 allows the argument to \c{TIMES} to contain expressions such as
1377 \c{64-$+buffer} as above. To repeat more than one line of code, or a
1378 complex macro, use the preprocessor \i\c{%rep} directive.
1381 \H{effaddr} Effective Addresses
1383 An \i{effective address} is any operand to an instruction which
1384 \I{memory reference}references memory. Effective addresses, in NASM,
1385 have a very simple syntax: they consist of an expression evaluating
1386 to the desired address, enclosed in \i{square brackets}. For
1391 \c mov ax,[wordvar+1]
1392 \c mov ax,[es:wordvar+bx]
1394 Anything not conforming to this simple system is not a valid memory
1395 reference in NASM, for example \c{es:wordvar[bx]}.
1397 More complicated effective addresses, such as those involving more
1398 than one register, work in exactly the same way:
1400 \c mov eax,[ebx*2+ecx+offset]
1403 NASM is capable of doing \i{algebra} on these effective addresses,
1404 so that things which don't necessarily \e{look} legal are perfectly
1407 \c mov eax,[ebx*5] ; assembles as [ebx*4+ebx]
1408 \c mov eax,[label1*2-label2] ; ie [label1+(label1-label2)]
1410 Some forms of effective address have more than one assembled form;
1411 in most such cases NASM will generate the smallest form it can. For
1412 example, there are distinct assembled forms for the 32-bit effective
1413 addresses \c{[eax*2+0]} and \c{[eax+eax]}, and NASM will generally
1414 generate the latter on the grounds that the former requires four
1415 bytes to store a zero offset.
1417 NASM has a hinting mechanism which will cause \c{[eax+ebx]} and
1418 \c{[ebx+eax]} to generate different opcodes; this is occasionally
1419 useful because \c{[esi+ebp]} and \c{[ebp+esi]} have different
1420 default segment registers.
1422 However, you can force NASM to generate an effective address in a
1423 particular form by the use of the keywords \c{BYTE}, \c{WORD},
1424 \c{DWORD} and \c{NOSPLIT}. If you need \c{[eax+3]} to be assembled
1425 using a double-word offset field instead of the one byte NASM will
1426 normally generate, you can code \c{[dword eax+3]}. Similarly, you
1427 can force NASM to use a byte offset for a small value which it
1428 hasn't seen on the first pass (see \k{crit} for an example of such a
1429 code fragment) by using \c{[byte eax+offset]}. As special cases,
1430 \c{[byte eax]} will code \c{[eax+0]} with a byte offset of zero, and
1431 \c{[dword eax]} will code it with a double-word offset of zero. The
1432 normal form, \c{[eax]}, will be coded with no offset field.
1434 The form described in the previous paragraph is also useful if you
1435 are trying to access data in a 32-bit segment from within 16 bit code.
1436 For more information on this see the section on mixed-size addressing
1437 (\k{mixaddr}). In particular, if you need to access data with a known
1438 offset that is larger than will fit in a 16-bit value, if you don't
1439 specify that it is a dword offset, nasm will cause the high word of
1440 the offset to be lost.
1442 Similarly, NASM will split \c{[eax*2]} into \c{[eax+eax]} because
1443 that allows the offset field to be absent and space to be saved; in
1444 fact, it will also split \c{[eax*2+offset]} into
1445 \c{[eax+eax+offset]}. You can combat this behaviour by the use of
1446 the \c{NOSPLIT} keyword: \c{[nosplit eax*2]} will force
1447 \c{[eax*2+0]} to be generated literally.
1449 In 64-bit mode, NASM will by default generate absolute addresses. The
1450 \i\c{REL} keyword makes it produce \c{RIP}-relative addresses. Since
1451 this is frequently the normally desired behaviour, see the \c{DEFAULT}
1452 directive (\k{default}). The keyword \i\c{ABS} overrides \i\c{REL}.
1455 \H{const} \i{Constants}
1457 NASM understands four different types of constant: numeric,
1458 character, string and floating-point.
1461 \S{numconst} \i{Numeric Constants}
1463 A numeric constant is simply a number. NASM allows you to specify
1464 numbers in a variety of number bases, in a variety of ways: you can
1465 suffix \c{H} or \c{X}, \c{D} or \c{T}, \c{Q} or \c{O}, and \c{B} or
1466 \c{Y} for \i{hexadecimal}, \i{decimal}, \i{octal} and \i{binary}
1467 respectively, or you can prefix \c{0x}, for hexadecimal in the style
1468 of C, or you can prefix \c{$} for hexadecimal in the style of Borland
1469 Pascal or Motorola Assemblers. Note, though, that the \I{$,
1470 prefix}\c{$} prefix does double duty as a prefix on identifiers (see
1471 \k{syntax}), so a hex number prefixed with a \c{$} sign must have a
1472 digit after the \c{$} rather than a letter. In addition, current
1473 versions of NASM accept the prefix \c{0h} for hexadecimal, \c{0d} or
1474 \c{0t} for decimal, \c{0o} or \c{0q} for octal, and \c{0b} or \c{0y}
1475 for binary. Please note that unlike C, a \c{0} prefix by itself does
1476 \e{not} imply an octal constant!
1478 Numeric constants can have underscores (\c{_}) interspersed to break
1481 Some examples (all producing exactly the same code):
1483 \c mov ax,200 ; decimal
1484 \c mov ax,0200 ; still decimal
1485 \c mov ax,0200d ; explicitly decimal
1486 \c mov ax,0d200 ; also decimal
1487 \c mov ax,0c8h ; hex
1488 \c mov ax,$0c8 ; hex again: the 0 is required
1489 \c mov ax,0xc8 ; hex yet again
1490 \c mov ax,0hc8 ; still hex
1491 \c mov ax,310q ; octal
1492 \c mov ax,310o ; octal again
1493 \c mov ax,0o310 ; octal yet again
1494 \c mov ax,0q310 ; octal yet again
1495 \c mov ax,11001000b ; binary
1496 \c mov ax,1100_1000b ; same binary constant
1497 \c mov ax,1100_1000y ; same binary constant once more
1498 \c mov ax,0b1100_1000 ; same binary constant yet again
1499 \c mov ax,0y1100_1000 ; same binary constant yet again
1501 \S{strings} \I{Strings}\i{Character Strings}
1503 A character string consists of up to eight characters enclosed in
1504 either single quotes (\c{'...'}), double quotes (\c{"..."}) or
1505 backquotes (\c{`...`}). Single or double quotes are equivalent to
1506 NASM (except of course that surrounding the constant with single
1507 quotes allows double quotes to appear within it and vice versa); the
1508 contents of those are represented verbatim. Strings enclosed in
1509 backquotes support C-style \c{\\}-escapes for special characters.
1512 The following \i{escape sequences} are recognized by backquoted strings:
1514 \c \' single quote (')
1515 \c \" double quote (")
1517 \c \\\ backslash (\)
1518 \c \? question mark (?)
1526 \c \e ESC (ASCII 27)
1527 \c \377 Up to 3 octal digits - literal byte
1528 \c \xFF Up to 2 hexadecimal digits - literal byte
1529 \c \u1234 4 hexadecimal digits - Unicode character
1530 \c \U12345678 8 hexadecimal digits - Unicode character
1532 All other escape sequences are reserved. Note that \c{\\0}, meaning a
1533 \c{NUL} character (ASCII 0), is a special case of the octal escape
1536 \i{Unicode} characters specified with \c{\\u} or \c{\\U} are converted to
1537 \i{UTF-8}. For example, the following lines are all equivalent:
1539 \c db `\u263a` ; UTF-8 smiley face
1540 \c db `\xe2\x98\xba` ; UTF-8 smiley face
1541 \c db 0E2h, 098h, 0BAh ; UTF-8 smiley face
1544 \S{chrconst} \i{Character Constants}
1546 A character constant consists of a string up to eight bytes long, used
1547 in an expression context. It is treated as if it was an integer.
1549 A character constant with more than one byte will be arranged
1550 with \i{little-endian} order in mind: if you code
1554 then the constant generated is not \c{0x61626364}, but
1555 \c{0x64636261}, so that if you were then to store the value into
1556 memory, it would read \c{abcd} rather than \c{dcba}. This is also
1557 the sense of character constants understood by the Pentium's
1558 \i\c{CPUID} instruction.
1561 \S{strconst} \i{String Constants}
1563 String constants are character strings used in the context of some
1564 pseudo-instructions, namely the
1565 \I\c{DW}\I\c{DD}\I\c{DQ}\I\c{DT}\I\c{DO}\I\c{DY}\i\c{DB} family and
1566 \i\c{INCBIN} (where it represents a filename.) They are also used in
1567 certain preprocessor directives.
1569 A string constant looks like a character constant, only longer. It
1570 is treated as a concatenation of maximum-size character constants
1571 for the conditions. So the following are equivalent:
1573 \c db 'hello' ; string constant
1574 \c db 'h','e','l','l','o' ; equivalent character constants
1576 And the following are also equivalent:
1578 \c dd 'ninechars' ; doubleword string constant
1579 \c dd 'nine','char','s' ; becomes three doublewords
1580 \c db 'ninechars',0,0,0 ; and really looks like this
1582 Note that when used in a string-supporting context, quoted strings are
1583 treated as a string constants even if they are short enough to be a
1584 character constant, because otherwise \c{db 'ab'} would have the same
1585 effect as \c{db 'a'}, which would be silly. Similarly, three-character
1586 or four-character constants are treated as strings when they are
1587 operands to \c{DW}, and so forth.
1589 \S{unicode} \I{UTF-16}\I{UTF-32}\i{Unicode} Strings
1591 The special operators \i\c{__utf16__} and \i\c{__utf32__} allows
1592 definition of Unicode strings. They take a string in UTF-8 format and
1593 converts it to (littleendian) UTF-16 or UTF-32, respectively.
1597 \c %define u(x) __utf16__(x)
1598 \c %define w(x) __utf32__(x)
1600 \c dw u('C:\WINDOWS'), 0 ; Pathname in UTF-16
1601 \c dd w(`A + B = \u206a`), 0 ; String in UTF-32
1603 \c{__utf16__} and \c{__utf32__} can be applied either to strings
1604 passed to the \c{DB} family instructions, or to character constants in
1605 an expression context.
1607 \S{fltconst} \I{floating-point, constants}Floating-Point Constants
1609 \i{Floating-point} constants are acceptable only as arguments to
1610 \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, and \i\c{DO}, or as
1611 arguments to the special operators \i\c{__float8__},
1612 \i\c{__float16__}, \i\c{__float32__}, \i\c{__float64__},
1613 \i\c{__float80m__}, \i\c{__float80e__}, \i\c{__float128l__}, and
1614 \i\c{__float128h__}.
1616 Floating-point constants are expressed in the traditional form:
1617 digits, then a period, then optionally more digits, then optionally an
1618 \c{E} followed by an exponent. The period is mandatory, so that NASM
1619 can distinguish between \c{dd 1}, which declares an integer constant,
1620 and \c{dd 1.0} which declares a floating-point constant.
1622 NASM also support C99-style hexadecimal floating-point: \c{0x},
1623 hexadecimal digits, period, optionally more hexadeximal digits, then
1624 optionally a \c{P} followed by a \e{binary} (not hexadecimal) exponent
1625 in decimal notation. As an extension, NASM additionally supports the
1626 \c{0h} and \c{$} prefixes for hexadecimal, as well binary and octal
1627 floating-point, using the \c{0b} or \c{0y} and \c{0o} or \c{0q}
1628 prefixes, respectively.
1630 Underscores to break up groups of digits are permitted in
1631 floating-point constants as well.
1635 \c db -0.2 ; "Quarter precision"
1636 \c dw -0.5 ; IEEE 754r/SSE5 half precision
1637 \c dd 1.2 ; an easy one
1638 \c dd 1.222_222_222 ; underscores are permitted
1639 \c dd 0x1p+2 ; 1.0x2^2 = 4.0
1640 \c dq 0x1p+32 ; 1.0x2^32 = 4 294 967 296.0
1641 \c dq 1.e10 ; 10 000 000 000.0
1642 \c dq 1.e+10 ; synonymous with 1.e10
1643 \c dq 1.e-10 ; 0.000 000 000 1
1644 \c dt 3.141592653589793238462 ; pi
1645 \c do 1.e+4000 ; IEEE 754r quad precision
1647 The 8-bit "quarter-precision" floating-point format is
1648 sign:exponent:mantissa = 1:4:3 with an exponent bias of 7. This
1649 appears to be the most frequently used 8-bit floating-point format,
1650 although it is not covered by any formal standard. This is sometimes
1651 called a "\i{minifloat}."
1653 The special operators are used to produce floating-point numbers in
1654 other contexts. They produce the binary representation of a specific
1655 floating-point number as an integer, and can use anywhere integer
1656 constants are used in an expression. \c{__float80m__} and
1657 \c{__float80e__} produce the 64-bit mantissa and 16-bit exponent of an
1658 80-bit floating-point number, and \c{__float128l__} and
1659 \c{__float128h__} produce the lower and upper 64-bit halves of a 128-bit
1660 floating-point number, respectively.
1664 \c mov rax,__float64__(3.141592653589793238462)
1666 ... would assign the binary representation of pi as a 64-bit floating
1667 point number into \c{RAX}. This is exactly equivalent to:
1669 \c mov rax,0x400921fb54442d18
1671 NASM cannot do compile-time arithmetic on floating-point constants.
1672 This is because NASM is designed to be portable - although it always
1673 generates code to run on x86 processors, the assembler itself can
1674 run on any system with an ANSI C compiler. Therefore, the assembler
1675 cannot guarantee the presence of a floating-point unit capable of
1676 handling the \i{Intel number formats}, and so for NASM to be able to
1677 do floating arithmetic it would have to include its own complete set
1678 of floating-point routines, which would significantly increase the
1679 size of the assembler for very little benefit.
1681 The special tokens \i\c{__Infinity__}, \i\c{__QNaN__} (or
1682 \i\c{__NaN__}) and \i\c{__SNaN__} can be used to generate
1683 \I{infinity}infinities, quiet \i{NaN}s, and signalling NaNs,
1684 respectively. These are normally used as macros:
1686 \c %define Inf __Infinity__
1687 \c %define NaN __QNaN__
1689 \c dq +1.5, -Inf, NaN ; Double-precision constants
1691 The \c{%use fp} standard macro package contains a set of convenience
1692 macros. See \k{pkg_fp}.
1694 \S{bcdconst} \I{floating-point, packed BCD constants}Packed BCD Constants
1696 x87-style packed BCD constants can be used in the same contexts as
1697 80-bit floating-point numbers. They are suffixed with \c{p} or
1698 prefixed with \c{0p}, and can include up to 18 decimal digits.
1700 As with other numeric constants, underscores can be used to separate
1705 \c dt 12_345_678_901_245_678p
1706 \c dt -12_345_678_901_245_678p
1711 \H{expr} \i{Expressions}
1713 Expressions in NASM are similar in syntax to those in C. Expressions
1714 are evaluated as 64-bit integers which are then adjusted to the
1717 NASM supports two special tokens in expressions, allowing
1718 calculations to involve the current assembly position: the
1719 \I{$, here}\c{$} and \i\c{$$} tokens. \c{$} evaluates to the assembly
1720 position at the beginning of the line containing the expression; so
1721 you can code an \i{infinite loop} using \c{JMP $}. \c{$$} evaluates
1722 to the beginning of the current section; so you can tell how far
1723 into the section you are by using \c{($-$$)}.
1725 The arithmetic \i{operators} provided by NASM are listed here, in
1726 increasing order of \i{precedence}.
1729 \S{expor} \i\c{|}: \i{Bitwise OR} Operator
1731 The \c{|} operator gives a bitwise OR, exactly as performed by the
1732 \c{OR} machine instruction. Bitwise OR is the lowest-priority
1733 arithmetic operator supported by NASM.
1736 \S{expxor} \i\c{^}: \i{Bitwise XOR} Operator
1738 \c{^} provides the bitwise XOR operation.
1741 \S{expand} \i\c{&}: \i{Bitwise AND} Operator
1743 \c{&} provides the bitwise AND operation.
1746 \S{expshift} \i\c{<<} and \i\c{>>}: \i{Bit Shift} Operators
1748 \c{<<} gives a bit-shift to the left, just as it does in C. So \c{5<<3}
1749 evaluates to 5 times 8, or 40. \c{>>} gives a bit-shift to the
1750 right; in NASM, such a shift is \e{always} unsigned, so that
1751 the bits shifted in from the left-hand end are filled with zero
1752 rather than a sign-extension of the previous highest bit.
1755 \S{expplmi} \I{+ opaddition}\c{+} and \I{- opsubtraction}\c{-}:
1756 \i{Addition} and \i{Subtraction} Operators
1758 The \c{+} and \c{-} operators do perfectly ordinary addition and
1762 \S{expmul} \i\c{*}, \i\c{/}, \i\c{//}, \i\c{%} and \i\c{%%}:
1763 \i{Multiplication} and \i{Division}
1765 \c{*} is the multiplication operator. \c{/} and \c{//} are both
1766 division operators: \c{/} is \i{unsigned division} and \c{//} is
1767 \i{signed division}. Similarly, \c{%} and \c{%%} provide \I{unsigned
1768 modulo}\I{modulo operators}unsigned and
1769 \i{signed modulo} operators respectively.
1771 NASM, like ANSI C, provides no guarantees about the sensible
1772 operation of the signed modulo operator.
1774 Since the \c{%} character is used extensively by the macro
1775 \i{preprocessor}, you should ensure that both the signed and unsigned
1776 modulo operators are followed by white space wherever they appear.
1779 \S{expmul} \i{Unary Operators}: \I{+ opunary}\c{+}, \I{- opunary}\c{-},
1780 \i\c{~}, \I{! opunary}\c{!} and \i\c{SEG}
1782 The highest-priority operators in NASM's expression grammar are
1783 those which only apply to one argument. \c{-} negates its operand,
1784 \c{+} does nothing (it's provided for symmetry with \c{-}), \c{~}
1785 computes the \i{one's complement} of its operand, \c{!} is the
1786 \i{logical negation} operator, and \c{SEG} provides the \i{segment address}
1787 of its operand (explained in more detail in \k{segwrt}).
1790 \H{segwrt} \i\c{SEG} and \i\c{WRT}
1792 When writing large 16-bit programs, which must be split into
1793 multiple \i{segments}, it is often necessary to be able to refer to
1794 the \I{segment address}segment part of the address of a symbol. NASM
1795 supports the \c{SEG} operator to perform this function.
1797 The \c{SEG} operator returns the \i\e{preferred} segment base of a
1798 symbol, defined as the segment base relative to which the offset of
1799 the symbol makes sense. So the code
1801 \c mov ax,seg symbol
1805 will load \c{ES:BX} with a valid pointer to the symbol \c{symbol}.
1807 Things can be more complex than this: since 16-bit segments and
1808 \i{groups} may \I{overlapping segments}overlap, you might occasionally
1809 want to refer to some symbol using a different segment base from the
1810 preferred one. NASM lets you do this, by the use of the \c{WRT}
1811 (With Reference To) keyword. So you can do things like
1813 \c mov ax,weird_seg ; weird_seg is a segment base
1815 \c mov bx,symbol wrt weird_seg
1817 to load \c{ES:BX} with a different, but functionally equivalent,
1818 pointer to the symbol \c{symbol}.
1820 NASM supports far (inter-segment) calls and jumps by means of the
1821 syntax \c{call segment:offset}, where \c{segment} and \c{offset}
1822 both represent immediate values. So to call a far procedure, you
1823 could code either of
1825 \c call (seg procedure):procedure
1826 \c call weird_seg:(procedure wrt weird_seg)
1828 (The parentheses are included for clarity, to show the intended
1829 parsing of the above instructions. They are not necessary in
1832 NASM supports the syntax \I\c{CALL FAR}\c{call far procedure} as a
1833 synonym for the first of the above usages. \c{JMP} works identically
1834 to \c{CALL} in these examples.
1836 To declare a \i{far pointer} to a data item in a data segment, you
1839 \c dw symbol, seg symbol
1841 NASM supports no convenient synonym for this, though you can always
1842 invent one using the macro processor.
1845 \H{strict} \i\c{STRICT}: Inhibiting Optimization
1847 When assembling with the optimizer set to level 2 or higher (see
1848 \k{opt-O}), NASM will use size specifiers (\c{BYTE}, \c{WORD},
1849 \c{DWORD}, \c{QWORD}, \c{TWORD}, \c{OWORD} or \c{YWORD}), but will
1850 give them the smallest possible size. The keyword \c{STRICT} can be
1851 used to inhibit optimization and force a particular operand to be
1852 emitted in the specified size. For example, with the optimizer on, and
1853 in \c{BITS 16} mode,
1857 is encoded in three bytes \c{66 6A 21}, whereas
1859 \c push strict dword 33
1861 is encoded in six bytes, with a full dword immediate operand \c{66 68
1864 With the optimizer off, the same code (six bytes) is generated whether
1865 the \c{STRICT} keyword was used or not.
1868 \H{crit} \i{Critical Expressions}
1870 Although NASM has an optional multi-pass optimizer, there are some
1871 expressions which must be resolvable on the first pass. These are
1872 called \e{Critical Expressions}.
1874 The first pass is used to determine the size of all the assembled
1875 code and data, so that the second pass, when generating all the
1876 code, knows all the symbol addresses the code refers to. So one
1877 thing NASM can't handle is code whose size depends on the value of a
1878 symbol declared after the code in question. For example,
1880 \c times (label-$) db 0
1881 \c label: db 'Where am I?'
1883 The argument to \i\c{TIMES} in this case could equally legally
1884 evaluate to anything at all; NASM will reject this example because
1885 it cannot tell the size of the \c{TIMES} line when it first sees it.
1886 It will just as firmly reject the slightly \I{paradox}paradoxical
1889 \c times (label-$+1) db 0
1890 \c label: db 'NOW where am I?'
1892 in which \e{any} value for the \c{TIMES} argument is by definition
1895 NASM rejects these examples by means of a concept called a
1896 \e{critical expression}, which is defined to be an expression whose
1897 value is required to be computable in the first pass, and which must
1898 therefore depend only on symbols defined before it. The argument to
1899 the \c{TIMES} prefix is a critical expression.
1901 \H{locallab} \i{Local Labels}
1903 NASM gives special treatment to symbols beginning with a \i{period}.
1904 A label beginning with a single period is treated as a \e{local}
1905 label, which means that it is associated with the previous non-local
1906 label. So, for example:
1908 \c label1 ; some code
1916 \c label2 ; some code
1924 In the above code fragment, each \c{JNE} instruction jumps to the
1925 line immediately before it, because the two definitions of \c{.loop}
1926 are kept separate by virtue of each being associated with the
1927 previous non-local label.
1929 This form of local label handling is borrowed from the old Amiga
1930 assembler \i{DevPac}; however, NASM goes one step further, in
1931 allowing access to local labels from other parts of the code. This
1932 is achieved by means of \e{defining} a local label in terms of the
1933 previous non-local label: the first definition of \c{.loop} above is
1934 really defining a symbol called \c{label1.loop}, and the second
1935 defines a symbol called \c{label2.loop}. So, if you really needed
1938 \c label3 ; some more code
1943 Sometimes it is useful - in a macro, for instance - to be able to
1944 define a label which can be referenced from anywhere but which
1945 doesn't interfere with the normal local-label mechanism. Such a
1946 label can't be non-local because it would interfere with subsequent
1947 definitions of, and references to, local labels; and it can't be
1948 local because the macro that defined it wouldn't know the label's
1949 full name. NASM therefore introduces a third type of label, which is
1950 probably only useful in macro definitions: if a label begins with
1951 the \I{label prefix}special prefix \i\c{..@}, then it does nothing
1952 to the local label mechanism. So you could code
1954 \c label1: ; a non-local label
1955 \c .local: ; this is really label1.local
1956 \c ..@foo: ; this is a special symbol
1957 \c label2: ; another non-local label
1958 \c .local: ; this is really label2.local
1960 \c jmp ..@foo ; this will jump three lines up
1962 NASM has the capacity to define other special symbols beginning with
1963 a double period: for example, \c{..start} is used to specify the
1964 entry point in the \c{obj} output format (see \k{dotdotstart}),
1965 \c{..imagebase} is used to find out the offset from a base address
1966 of the current image in the \c{win64} output format (see \k{win64pic}).
1967 So just keep in mind that symbols beginning with a double period are
1971 \C{preproc} The NASM \i{Preprocessor}
1973 NASM contains a powerful \i{macro processor}, which supports
1974 conditional assembly, multi-level file inclusion, two forms of macro
1975 (single-line and multi-line), and a `context stack' mechanism for
1976 extra macro power. Preprocessor directives all begin with a \c{%}
1979 The preprocessor collapses all lines which end with a backslash (\\)
1980 character into a single line. Thus:
1982 \c %define THIS_VERY_LONG_MACRO_NAME_IS_DEFINED_TO \\
1985 will work like a single-line macro without the backslash-newline
1988 \H{slmacro} \i{Single-Line Macros}
1990 \S{define} The Normal Way: \I\c{%idefine}\i\c{%define}
1992 Single-line macros are defined using the \c{%define} preprocessor
1993 directive. The definitions work in a similar way to C; so you can do
1996 \c %define ctrl 0x1F &
1997 \c %define param(a,b) ((a)+(a)*(b))
1999 \c mov byte [param(2,ebx)], ctrl 'D'
2001 which will expand to
2003 \c mov byte [(2)+(2)*(ebx)], 0x1F & 'D'
2005 When the expansion of a single-line macro contains tokens which
2006 invoke another macro, the expansion is performed at invocation time,
2007 not at definition time. Thus the code
2009 \c %define a(x) 1+b(x)
2014 will evaluate in the expected way to \c{mov ax,1+2*8}, even though
2015 the macro \c{b} wasn't defined at the time of definition of \c{a}.
2017 Macros defined with \c{%define} are \i{case sensitive}: after
2018 \c{%define foo bar}, only \c{foo} will expand to \c{bar}: \c{Foo} or
2019 \c{FOO} will not. By using \c{%idefine} instead of \c{%define} (the
2020 `i' stands for `insensitive') you can define all the case variants
2021 of a macro at once, so that \c{%idefine foo bar} would cause
2022 \c{foo}, \c{Foo}, \c{FOO}, \c{fOO} and so on all to expand to
2025 There is a mechanism which detects when a macro call has occurred as
2026 a result of a previous expansion of the same macro, to guard against
2027 \i{circular references} and infinite loops. If this happens, the
2028 preprocessor will only expand the first occurrence of the macro.
2031 \c %define a(x) 1+a(x)
2035 the macro \c{a(3)} will expand once, becoming \c{1+a(3)}, and will
2036 then expand no further. This behaviour can be useful: see \k{32c}
2037 for an example of its use.
2039 You can \I{overloading, single-line macros}overload single-line
2040 macros: if you write
2042 \c %define foo(x) 1+x
2043 \c %define foo(x,y) 1+x*y
2045 the preprocessor will be able to handle both types of macro call,
2046 by counting the parameters you pass; so \c{foo(3)} will become
2047 \c{1+3} whereas \c{foo(ebx,2)} will become \c{1+ebx*2}. However, if
2052 then no other definition of \c{foo} will be accepted: a macro with
2053 no parameters prohibits the definition of the same name as a macro
2054 \e{with} parameters, and vice versa.
2056 This doesn't prevent single-line macros being \e{redefined}: you can
2057 perfectly well define a macro with
2061 and then re-define it later in the same source file with
2065 Then everywhere the macro \c{foo} is invoked, it will be expanded
2066 according to the most recent definition. This is particularly useful
2067 when defining single-line macros with \c{%assign} (see \k{assign}).
2069 You can \i{pre-define} single-line macros using the `-d' option on
2070 the NASM command line: see \k{opt-d}.
2073 \S{xdefine} Resolving \c{%define}: \I\c{%ixdefine}\i\c{%xdefine}
2075 To have a reference to an embedded single-line macro resolved at the
2076 time that the embedding macro is \e{defined}, as opposed to when the
2077 embedding macro is \e{expanded}, you need a different mechanism to the
2078 one offered by \c{%define}. The solution is to use \c{%xdefine}, or
2079 it's \I{case sensitive}case-insensitive counterpart \c{%ixdefine}.
2081 Suppose you have the following code:
2084 \c %define isFalse isTrue
2093 In this case, \c{val1} is equal to 0, and \c{val2} is equal to 1.
2094 This is because, when a single-line macro is defined using
2095 \c{%define}, it is expanded only when it is called. As \c{isFalse}
2096 expands to \c{isTrue}, the expansion will be the current value of
2097 \c{isTrue}. The first time it is called that is 0, and the second
2100 If you wanted \c{isFalse} to expand to the value assigned to the
2101 embedded macro \c{isTrue} at the time that \c{isFalse} was defined,
2102 you need to change the above code to use \c{%xdefine}.
2104 \c %xdefine isTrue 1
2105 \c %xdefine isFalse isTrue
2106 \c %xdefine isTrue 0
2110 \c %xdefine isTrue 1
2114 Now, each time that \c{isFalse} is called, it expands to 1,
2115 as that is what the embedded macro \c{isTrue} expanded to at
2116 the time that \c{isFalse} was defined.
2119 \S{indmacro} \i{Macro Indirection}: \I\c{%[}\c{%[...]}
2121 The \c{%[...]} construct can be used to expand macros in contexts
2122 where macro expansion would otherwise not occur, including in the
2123 names other macros. For example, if you have a set of macros named
2124 \c{Foo16}, \c{Foo32} and \c{Foo64}, you could write:
2126 \c mov ax,Foo%[__BITS__] ; The Foo value
2128 to use the builtin macro \c{__BITS__} (see \k{bitsm}) to automatically
2129 select between them. Similarly, the two statements:
2131 \c %xdefine Bar Quux ; Expands due to %xdefine
2132 \c %define Bar %[Quux] ; Expands due to %[...]
2134 have, in fact, exactly the same effect.
2136 \c{%[...]} concatenates to adjacent tokens in the same way that
2137 multi-line macro parameters do, see \k{concat} for details.
2140 \S{concat%+} Concatenating Single Line Macro Tokens: \i\c{%+}
2142 Individual tokens in single line macros can be concatenated, to produce
2143 longer tokens for later processing. This can be useful if there are
2144 several similar macros that perform similar functions.
2146 Please note that a space is required after \c{%+}, in order to
2147 disambiguate it from the syntax \c{%+1} used in multiline macros.
2149 As an example, consider the following:
2151 \c %define BDASTART 400h ; Start of BIOS data area
2153 \c struc tBIOSDA ; its structure
2159 Now, if we need to access the elements of tBIOSDA in different places,
2162 \c mov ax,BDASTART + tBIOSDA.COM1addr
2163 \c mov bx,BDASTART + tBIOSDA.COM2addr
2165 This will become pretty ugly (and tedious) if used in many places, and
2166 can be reduced in size significantly by using the following macro:
2168 \c ; Macro to access BIOS variables by their names (from tBDA):
2170 \c %define BDA(x) BDASTART + tBIOSDA. %+ x
2172 Now the above code can be written as:
2174 \c mov ax,BDA(COM1addr)
2175 \c mov bx,BDA(COM2addr)
2177 Using this feature, we can simplify references to a lot of macros (and,
2178 in turn, reduce typing errors).
2181 \S{selfref%?} The Macro Name Itself: \i\c{%?} and \i\c{%??}
2183 The special symbols \c{%?} and \c{%??} can be used to reference the
2184 macro name itself inside a macro expansion, this is supported for both
2185 single-and multi-line macros. \c{%?} refers to the macro name as
2186 \e{invoked}, whereas \c{%??} refers to the macro name as
2187 \e{declared}. The two are always the same for case-sensitive
2188 macros, but for case-insensitive macros, they can differ.
2192 \c %idefine Foo mov %?,%??
2204 \c %idefine keyword $%?
2206 can be used to make a keyword "disappear", for example in case a new
2207 instruction has been used as a label in older code. For example:
2209 \c %idefine pause $%? ; Hide the PAUSE instruction
2212 \S{undef} Undefining Single-Line Macros: \i\c{%undef}
2214 Single-line macros can be removed with the \c{%undef} directive. For
2215 example, the following sequence:
2222 will expand to the instruction \c{mov eax, foo}, since after
2223 \c{%undef} the macro \c{foo} is no longer defined.
2225 Macros that would otherwise be pre-defined can be undefined on the
2226 command-line using the `-u' option on the NASM command line: see
2230 \S{assign} \i{Preprocessor Variables}: \i\c{%assign}
2232 An alternative way to define single-line macros is by means of the
2233 \c{%assign} command (and its \I{case sensitive}case-insensitive
2234 counterpart \i\c{%iassign}, which differs from \c{%assign} in
2235 exactly the same way that \c{%idefine} differs from \c{%define}).
2237 \c{%assign} is used to define single-line macros which take no
2238 parameters and have a numeric value. This value can be specified in
2239 the form of an expression, and it will be evaluated once, when the
2240 \c{%assign} directive is processed.
2242 Like \c{%define}, macros defined using \c{%assign} can be re-defined
2243 later, so you can do things like
2247 to increment the numeric value of a macro.
2249 \c{%assign} is useful for controlling the termination of \c{%rep}
2250 preprocessor loops: see \k{rep} for an example of this. Another
2251 use for \c{%assign} is given in \k{16c} and \k{32c}.
2253 The expression passed to \c{%assign} is a \i{critical expression}
2254 (see \k{crit}), and must also evaluate to a pure number (rather than
2255 a relocatable reference such as a code or data address, or anything
2256 involving a register).
2259 \S{defstr} Defining Strings: \I\c{%idefstr}\i\c{%defstr}
2261 \c{%defstr}, and its case-insensitive counterpart \c{%idefstr}, define
2262 or redefine a single-line macro without parameters but converts the
2263 entire right-hand side, after macro expansion, to a quoted string
2268 \c %defstr test TEST
2272 \c %define test 'TEST'
2274 This can be used, for example, with the \c{%!} construct (see
2277 \c %defstr PATH %!PATH ; The operating system PATH variable
2280 \S{deftok} Defining Tokens: \I\c{%ideftok}\i\c{%deftok}
2282 \c{%deftok}, and its case-insensitive counterpart \c{%ideftok}, define
2283 or redefine a single-line macro without parameters but converts the
2284 second parameter, after string conversion, to a sequence of tokens.
2288 \c %deftok test 'TEST'
2292 \c %define test TEST
2295 \H{strlen} \i{String Manipulation in Macros}
2297 It's often useful to be able to handle strings in macros. NASM
2298 supports a few simple string handling macro operators from which
2299 more complex operations can be constructed.
2301 All the string operators define or redefine a value (either a string
2302 or a numeric value) to a single-line macro. When producing a string
2303 value, it may change the style of quoting of the input string or
2304 strings, and possibly use \c{\\}-escapes inside \c{`}-quoted strings.
2306 \S{strcat} \i{Concatenating Strings}: \i\c{%strcat}
2308 The \c{%strcat} operator concatenates quoted strings and assign them to
2309 a single-line macro.
2313 \c %strcat alpha "Alpha: ", '12" screen'
2315 ... would assign the value \c{'Alpha: 12" screen'} to \c{alpha}.
2318 \c %strcat beta '"foo"\', "'bar'"
2320 ... would assign the value \c{`"foo"\\\\'bar'`} to \c{beta}.
2322 The use of commas to separate strings is permitted but optional.
2325 \S{strlen} \i{String Length}: \i\c{%strlen}
2327 The \c{%strlen} operator assigns the length of a string to a macro.
2330 \c %strlen charcnt 'my string'
2332 In this example, \c{charcnt} would receive the value 9, just as
2333 if an \c{%assign} had been used. In this example, \c{'my string'}
2334 was a literal string but it could also have been a single-line
2335 macro that expands to a string, as in the following example:
2337 \c %define sometext 'my string'
2338 \c %strlen charcnt sometext
2340 As in the first case, this would result in \c{charcnt} being
2341 assigned the value of 9.
2344 \S{substr} \i{Extracting Substrings}: \i\c{%substr}
2346 Individual letters or substrings in strings can be extracted using the
2347 \c{%substr} operator. An example of its use is probably more useful
2348 than the description:
2350 \c %substr mychar 'xyzw' 1 ; equivalent to %define mychar 'x'
2351 \c %substr mychar 'xyzw' 2 ; equivalent to %define mychar 'y'
2352 \c %substr mychar 'xyzw' 3 ; equivalent to %define mychar 'z'
2353 \c %substr mychar 'xyzw' 2,2 ; equivalent to %define mychar 'yz'
2354 \c %substr mychar 'xyzw' 2,-1 ; equivalent to %define mychar 'yzw'
2355 \c %substr mychar 'xyzw' 2,-2 ; equivalent to %define mychar 'yz'
2357 As with \c{%strlen} (see \k{strlen}), the first parameter is the
2358 single-line macro to be created and the second is the string. The
2359 third parameter specifies the first character to be selected, and the
2360 optional fourth parameter preceeded by comma) is the length. Note
2361 that the first index is 1, not 0 and the last index is equal to the
2362 value that \c{%strlen} would assign given the same string. Index
2363 values out of range result in an empty string. A negative length
2364 means "until N-1 characters before the end of string", i.e. \c{-1}
2365 means until end of string, \c{-2} until one character before, etc.
2368 \H{mlmacro} \i{Multi-Line Macros}: \I\c{%imacro}\i\c{%macro}
2370 Multi-line macros are much more like the type of macro seen in MASM
2371 and TASM: a multi-line macro definition in NASM looks something like
2374 \c %macro prologue 1
2382 This defines a C-like function prologue as a macro: so you would
2383 invoke the macro with a call such as
2385 \c myfunc: prologue 12
2387 which would expand to the three lines of code
2393 The number \c{1} after the macro name in the \c{%macro} line defines
2394 the number of parameters the macro \c{prologue} expects to receive.
2395 The use of \c{%1} inside the macro definition refers to the first
2396 parameter to the macro call. With a macro taking more than one
2397 parameter, subsequent parameters would be referred to as \c{%2},
2400 Multi-line macros, like single-line macros, are \i{case-sensitive},
2401 unless you define them using the alternative directive \c{%imacro}.
2403 If you need to pass a comma as \e{part} of a parameter to a
2404 multi-line macro, you can do that by enclosing the entire parameter
2405 in \I{braces, around macro parameters}braces. So you could code
2414 \c silly 'a', letter_a ; letter_a: db 'a'
2415 \c silly 'ab', string_ab ; string_ab: db 'ab'
2416 \c silly {13,10}, crlf ; crlf: db 13,10
2419 \S{mlrmacro} \i{Recursive Multi-Line Macros}: \I\c{%irmacro}\i\c{%rmacro}
2421 A multi-line macro cannot be referenced within itself, in order to
2422 prevent accidental infinite recursion and allow instruction overloading.
2424 Recursive multi-line macros allow for self-referencing, with the
2425 caveat that the user is aware of the existence, use and purpose of
2426 recursive multi-line macros. There is also a generous, but sane, upper
2427 limit to the number of recursions, in order to prevent run-away memory
2428 consumption in case of accidental infinite recursion.
2430 As with non-recursive multi-line macros, recursive multi-line macros are
2431 \i{case-sensitive}, unless you define them using the alternative
2432 directive \c{%irmacro}.
2435 \S{mlmacover} Overloading Multi-Line Macros\I{overloading, multi-line macros}
2437 As with single-line macros, multi-line macros can be overloaded by
2438 defining the same macro name several times with different numbers of
2439 parameters. This time, no exception is made for macros with no
2440 parameters at all. So you could define
2442 \c %macro prologue 0
2449 to define an alternative form of the function prologue which
2450 allocates no local stack space.
2452 Sometimes, however, you might want to `overload' a machine
2453 instruction; for example, you might want to define
2462 so that you could code
2464 \c push ebx ; this line is not a macro call
2465 \c push eax,ecx ; but this one is
2467 Ordinarily, NASM will give a warning for the first of the above two
2468 lines, since \c{push} is now defined to be a macro, and is being
2469 invoked with a number of parameters for which no definition has been
2470 given. The correct code will still be generated, but the assembler
2471 will give a warning. This warning can be disabled by the use of the
2472 \c{-w-macro-params} command-line option (see \k{opt-w}).
2475 \S{maclocal} \i{Macro-Local Labels}
2477 NASM allows you to define labels within a multi-line macro
2478 definition in such a way as to make them local to the macro call: so
2479 calling the same macro multiple times will use a different label
2480 each time. You do this by prefixing \i\c{%%} to the label name. So
2481 you can invent an instruction which executes a \c{RET} if the \c{Z}
2482 flag is set by doing this:
2492 You can call this macro as many times as you want, and every time
2493 you call it NASM will make up a different `real' name to substitute
2494 for the label \c{%%skip}. The names NASM invents are of the form
2495 \c{..@2345.skip}, where the number 2345 changes with every macro
2496 call. The \i\c{..@} prefix prevents macro-local labels from
2497 interfering with the local label mechanism, as described in
2498 \k{locallab}. You should avoid defining your own labels in this form
2499 (the \c{..@} prefix, then a number, then another period) in case
2500 they interfere with macro-local labels.
2503 \S{mlmacgre} \i{Greedy Macro Parameters}
2505 Occasionally it is useful to define a macro which lumps its entire
2506 command line into one parameter definition, possibly after
2507 extracting one or two smaller parameters from the front. An example
2508 might be a macro to write a text string to a file in MS-DOS, where
2509 you might want to be able to write
2511 \c writefile [filehandle],"hello, world",13,10
2513 NASM allows you to define the last parameter of a macro to be
2514 \e{greedy}, meaning that if you invoke the macro with more
2515 parameters than it expects, all the spare parameters get lumped into
2516 the last defined one along with the separating commas. So if you
2519 \c %macro writefile 2+
2525 \c mov cx,%%endstr-%%str
2532 then the example call to \c{writefile} above will work as expected:
2533 the text before the first comma, \c{[filehandle]}, is used as the
2534 first macro parameter and expanded when \c{%1} is referred to, and
2535 all the subsequent text is lumped into \c{%2} and placed after the
2538 The greedy nature of the macro is indicated to NASM by the use of
2539 the \I{+ modifier}\c{+} sign after the parameter count on the
2542 If you define a greedy macro, you are effectively telling NASM how
2543 it should expand the macro given \e{any} number of parameters from
2544 the actual number specified up to infinity; in this case, for
2545 example, NASM now knows what to do when it sees a call to
2546 \c{writefile} with 2, 3, 4 or more parameters. NASM will take this
2547 into account when overloading macros, and will not allow you to
2548 define another form of \c{writefile} taking 4 parameters (for
2551 Of course, the above macro could have been implemented as a
2552 non-greedy macro, in which case the call to it would have had to
2555 \c writefile [filehandle], {"hello, world",13,10}
2557 NASM provides both mechanisms for putting \i{commas in macro
2558 parameters}, and you choose which one you prefer for each macro
2561 See \k{sectmac} for a better way to write the above macro.
2563 \S{mlmacrange} \i{Macro Parameters Range}
2565 NASM allows you to expand parameters via special construction \c{%\{x:y\}}
2566 where \c{x} is the first parameter index and \c{y} is the last. Any index can
2567 be either negative or positive but must never be zero.
2577 expands to \c{3,4,5} range.
2579 Even more, the parameters can be reversed so that
2587 expands to \c{5,4,3} range.
2589 But even this is not the last. The parameters can be addressed via negative
2590 indices so NASM will count them reversed. The ones who know Python may see
2599 expands to \c{6,5,4} range.
2601 Note that NASM uses \i{comma} to separate parameters being expanded.
2603 By the way, here is a trick - you might use the index \c{%{-1:-1}}
2604 which gives you the \i{last} argument passed to a macro.
2606 \S{mlmacdef} \i{Default Macro Parameters}
2608 NASM also allows you to define a multi-line macro with a \e{range}
2609 of allowable parameter counts. If you do this, you can specify
2610 defaults for \i{omitted parameters}. So, for example:
2612 \c %macro die 0-1 "Painful program death has occurred."
2620 This macro (which makes use of the \c{writefile} macro defined in
2621 \k{mlmacgre}) can be called with an explicit error message, which it
2622 will display on the error output stream before exiting, or it can be
2623 called with no parameters, in which case it will use the default
2624 error message supplied in the macro definition.
2626 In general, you supply a minimum and maximum number of parameters
2627 for a macro of this type; the minimum number of parameters are then
2628 required in the macro call, and then you provide defaults for the
2629 optional ones. So if a macro definition began with the line
2631 \c %macro foobar 1-3 eax,[ebx+2]
2633 then it could be called with between one and three parameters, and
2634 \c{%1} would always be taken from the macro call. \c{%2}, if not
2635 specified by the macro call, would default to \c{eax}, and \c{%3} if
2636 not specified would default to \c{[ebx+2]}.
2638 You can provide extra information to a macro by providing
2639 too many default parameters:
2641 \c %macro quux 1 something
2643 This will trigger a warning by default; see \k{opt-w} for
2645 When \c{quux} is invoked, it receives not one but two parameters.
2646 \c{something} can be referred to as \c{%2}. The difference
2647 between passing \c{something} this way and writing \c{something}
2648 in the macro body is that with this way \c{something} is evaluated
2649 when the macro is defined, not when it is expanded.
2651 You may omit parameter defaults from the macro definition, in which
2652 case the parameter default is taken to be blank. This can be useful
2653 for macros which can take a variable number of parameters, since the
2654 \i\c{%0} token (see \k{percent0}) allows you to determine how many
2655 parameters were really passed to the macro call.
2657 This defaulting mechanism can be combined with the greedy-parameter
2658 mechanism; so the \c{die} macro above could be made more powerful,
2659 and more useful, by changing the first line of the definition to
2661 \c %macro die 0-1+ "Painful program death has occurred.",13,10
2663 The maximum parameter count can be infinite, denoted by \c{*}. In
2664 this case, of course, it is impossible to provide a \e{full} set of
2665 default parameters. Examples of this usage are shown in \k{rotate}.
2668 \S{percent0} \i\c{%0}: \I{counting macro parameters}Macro Parameter Counter
2670 The parameter reference \c{%0} will return a numeric constant giving the
2671 number of parameters received, that is, if \c{%0} is n then \c{%}n is the
2672 last parameter. \c{%0} is mostly useful for macros that can take a variable
2673 number of parameters. It can be used as an argument to \c{%rep}
2674 (see \k{rep}) in order to iterate through all the parameters of a macro.
2675 Examples are given in \k{rotate}.
2678 \S{percent00} \i\c{%00}: \I{label preceeding macro}Label Preceeding Macro
2680 \c{%00} will return the label preceeding the macro invocation, if any. The
2681 label must be on the same line as the macro invocation, may be a local label
2682 (see \k{locallab}), and need not end in a colon.
2685 \S{rotate} \i\c{%rotate}: \i{Rotating Macro Parameters}
2687 Unix shell programmers will be familiar with the \I{shift
2688 command}\c{shift} shell command, which allows the arguments passed
2689 to a shell script (referenced as \c{$1}, \c{$2} and so on) to be
2690 moved left by one place, so that the argument previously referenced
2691 as \c{$2} becomes available as \c{$1}, and the argument previously
2692 referenced as \c{$1} is no longer available at all.
2694 NASM provides a similar mechanism, in the form of \c{%rotate}. As
2695 its name suggests, it differs from the Unix \c{shift} in that no
2696 parameters are lost: parameters rotated off the left end of the
2697 argument list reappear on the right, and vice versa.
2699 \c{%rotate} is invoked with a single numeric argument (which may be
2700 an expression). The macro parameters are rotated to the left by that
2701 many places. If the argument to \c{%rotate} is negative, the macro
2702 parameters are rotated to the right.
2704 \I{iterating over macro parameters}So a pair of macros to save and
2705 restore a set of registers might work as follows:
2707 \c %macro multipush 1-*
2716 This macro invokes the \c{PUSH} instruction on each of its arguments
2717 in turn, from left to right. It begins by pushing its first
2718 argument, \c{%1}, then invokes \c{%rotate} to move all the arguments
2719 one place to the left, so that the original second argument is now
2720 available as \c{%1}. Repeating this procedure as many times as there
2721 were arguments (achieved by supplying \c{%0} as the argument to
2722 \c{%rep}) causes each argument in turn to be pushed.
2724 Note also the use of \c{*} as the maximum parameter count,
2725 indicating that there is no upper limit on the number of parameters
2726 you may supply to the \i\c{multipush} macro.
2728 It would be convenient, when using this macro, to have a \c{POP}
2729 equivalent, which \e{didn't} require the arguments to be given in
2730 reverse order. Ideally, you would write the \c{multipush} macro
2731 call, then cut-and-paste the line to where the pop needed to be
2732 done, and change the name of the called macro to \c{multipop}, and
2733 the macro would take care of popping the registers in the opposite
2734 order from the one in which they were pushed.
2736 This can be done by the following definition:
2738 \c %macro multipop 1-*
2747 This macro begins by rotating its arguments one place to the
2748 \e{right}, so that the original \e{last} argument appears as \c{%1}.
2749 This is then popped, and the arguments are rotated right again, so
2750 the second-to-last argument becomes \c{%1}. Thus the arguments are
2751 iterated through in reverse order.
2754 \S{concat} \i{Concatenating Macro Parameters}
2756 NASM can concatenate macro parameters and macro indirection constructs
2757 on to other text surrounding them. This allows you to declare a family
2758 of symbols, for example, in a macro definition. If, for example, you
2759 wanted to generate a table of key codes along with offsets into the
2760 table, you could code something like
2762 \c %macro keytab_entry 2
2764 \c keypos%1 equ $-keytab
2770 \c keytab_entry F1,128+1
2771 \c keytab_entry F2,128+2
2772 \c keytab_entry Return,13
2774 which would expand to
2777 \c keyposF1 equ $-keytab
2779 \c keyposF2 equ $-keytab
2781 \c keyposReturn equ $-keytab
2784 You can just as easily concatenate text on to the other end of a
2785 macro parameter, by writing \c{%1foo}.
2787 If you need to append a \e{digit} to a macro parameter, for example
2788 defining labels \c{foo1} and \c{foo2} when passed the parameter
2789 \c{foo}, you can't code \c{%11} because that would be taken as the
2790 eleventh macro parameter. Instead, you must code
2791 \I{braces, after % sign}\c{%\{1\}1}, which will separate the first
2792 \c{1} (giving the number of the macro parameter) from the second
2793 (literal text to be concatenated to the parameter).
2795 This concatenation can also be applied to other preprocessor in-line
2796 objects, such as macro-local labels (\k{maclocal}) and context-local
2797 labels (\k{ctxlocal}). In all cases, ambiguities in syntax can be
2798 resolved by enclosing everything after the \c{%} sign and before the
2799 literal text in braces: so \c{%\{%foo\}bar} concatenates the text
2800 \c{bar} to the end of the real name of the macro-local label
2801 \c{%%foo}. (This is unnecessary, since the form NASM uses for the
2802 real names of macro-local labels means that the two usages
2803 \c{%\{%foo\}bar} and \c{%%foobar} would both expand to the same
2804 thing anyway; nevertheless, the capability is there.)
2806 The single-line macro indirection construct, \c{%[...]}
2807 (\k{indmacro}), behaves the same way as macro parameters for the
2808 purpose of concatenation.
2810 See also the \c{%+} operator, \k{concat%+}.
2813 \S{mlmaccc} \i{Condition Codes as Macro Parameters}
2815 NASM can give special treatment to a macro parameter which contains
2816 a condition code. For a start, you can refer to the macro parameter
2817 \c{%1} by means of the alternative syntax \i\c{%+1}, which informs
2818 NASM that this macro parameter is supposed to contain a condition
2819 code, and will cause the preprocessor to report an error message if
2820 the macro is called with a parameter which is \e{not} a valid
2823 Far more usefully, though, you can refer to the macro parameter by
2824 means of \i\c{%-1}, which NASM will expand as the \e{inverse}
2825 condition code. So the \c{retz} macro defined in \k{maclocal} can be
2826 replaced by a general \i{conditional-return macro} like this:
2836 This macro can now be invoked using calls like \c{retc ne}, which
2837 will cause the conditional-jump instruction in the macro expansion
2838 to come out as \c{JE}, or \c{retc po} which will make the jump a
2841 The \c{%+1} macro-parameter reference is quite happy to interpret
2842 the arguments \c{CXZ} and \c{ECXZ} as valid condition codes;
2843 however, \c{%-1} will report an error if passed either of these,
2844 because no inverse condition code exists.
2847 \S{nolist} \i{Disabling Listing Expansion}\I\c{.nolist}
2849 When NASM is generating a listing file from your program, it will
2850 generally expand multi-line macros by means of writing the macro
2851 call and then listing each line of the expansion. This allows you to
2852 see which instructions in the macro expansion are generating what
2853 code; however, for some macros this clutters the listing up
2856 NASM therefore provides the \c{.nolist} qualifier, which you can
2857 include in a macro definition to inhibit the expansion of the macro
2858 in the listing file. The \c{.nolist} qualifier comes directly after
2859 the number of parameters, like this:
2861 \c %macro foo 1.nolist
2865 \c %macro bar 1-5+.nolist a,b,c,d,e,f,g,h
2867 \S{unmacro} Undefining Multi-Line Macros: \i\c{%unmacro}
2869 Multi-line macros can be removed with the \c{%unmacro} directive.
2870 Unlike the \c{%undef} directive, however, \c{%unmacro} takes an
2871 argument specification, and will only remove \i{exact matches} with
2872 that argument specification.
2881 removes the previously defined macro \c{foo}, but
2888 does \e{not} remove the macro \c{bar}, since the argument
2889 specification does not match exactly.
2892 \S{exitmacro} Exiting Multi-Line Macros: \i\c{%exitmacro}
2894 Multi-line macro expansions can be arbitrarily terminated with
2895 the \c{%exitmacro} directive.
2908 \H{condasm} \i{Conditional Assembly}\I\c{%if}
2910 Similarly to the C preprocessor, NASM allows sections of a source
2911 file to be assembled only if certain conditions are met. The general
2912 syntax of this feature looks like this:
2915 \c ; some code which only appears if <condition> is met
2916 \c %elif<condition2>
2917 \c ; only appears if <condition> is not met but <condition2> is
2919 \c ; this appears if neither <condition> nor <condition2> was met
2922 The inverse forms \i\c{%ifn} and \i\c{%elifn} are also supported.
2924 The \i\c{%else} clause is optional, as is the \i\c{%elif} clause.
2925 You can have more than one \c{%elif} clause as well.
2927 There are a number of variants of the \c{%if} directive. Each has its
2928 corresponding \c{%elif}, \c{%ifn}, and \c{%elifn} directives; for
2929 example, the equivalents to the \c{%ifdef} directive are \c{%elifdef},
2930 \c{%ifndef}, and \c{%elifndef}.
2932 \S{ifdef} \i\c{%ifdef}: Testing Single-Line Macro Existence\I{testing,
2933 single-line macro existence}
2935 Beginning a conditional-assembly block with the line \c{%ifdef
2936 MACRO} will assemble the subsequent code if, and only if, a
2937 single-line macro called \c{MACRO} is defined. If not, then the
2938 \c{%elif} and \c{%else} blocks (if any) will be processed instead.
2940 For example, when debugging a program, you might want to write code
2943 \c ; perform some function
2945 \c writefile 2,"Function performed successfully",13,10
2947 \c ; go and do something else
2949 Then you could use the command-line option \c{-dDEBUG} to create a
2950 version of the program which produced debugging messages, and remove
2951 the option to generate the final release version of the program.
2953 You can test for a macro \e{not} being defined by using
2954 \i\c{%ifndef} instead of \c{%ifdef}. You can also test for macro
2955 definitions in \c{%elif} blocks by using \i\c{%elifdef} and
2959 \S{ifmacro} \i\c{%ifmacro}: Testing Multi-Line Macro
2960 Existence\I{testing, multi-line macro existence}
2962 The \c{%ifmacro} directive operates in the same way as the \c{%ifdef}
2963 directive, except that it checks for the existence of a multi-line macro.
2965 For example, you may be working with a large project and not have control
2966 over the macros in a library. You may want to create a macro with one
2967 name if it doesn't already exist, and another name if one with that name
2970 The \c{%ifmacro} is considered true if defining a macro with the given name
2971 and number of arguments would cause a definitions conflict. For example:
2973 \c %ifmacro MyMacro 1-3
2975 \c %error "MyMacro 1-3" causes a conflict with an existing macro.
2979 \c %macro MyMacro 1-3
2981 \c ; insert code to define the macro
2987 This will create the macro "MyMacro 1-3" if no macro already exists which
2988 would conflict with it, and emits a warning if there would be a definition
2991 You can test for the macro not existing by using the \i\c{%ifnmacro} instead
2992 of \c{%ifmacro}. Additional tests can be performed in \c{%elif} blocks by using
2993 \i\c{%elifmacro} and \i\c{%elifnmacro}.
2996 \S{ifctx} \i\c{%ifctx}: Testing the Context Stack\I{testing, context
2999 The conditional-assembly construct \c{%ifctx} will cause the
3000 subsequent code to be assembled if and only if the top context on
3001 the preprocessor's context stack has the same name as one of the arguments.
3002 As with \c{%ifdef}, the inverse and \c{%elif} forms \i\c{%ifnctx},
3003 \i\c{%elifctx} and \i\c{%elifnctx} are also supported.
3005 For more details of the context stack, see \k{ctxstack}. For a
3006 sample use of \c{%ifctx}, see \k{blockif}.
3009 \S{if} \i\c{%if}: Testing Arbitrary Numeric Expressions\I{testing,
3010 arbitrary numeric expressions}
3012 The conditional-assembly construct \c{%if expr} will cause the
3013 subsequent code to be assembled if and only if the value of the
3014 numeric expression \c{expr} is non-zero. An example of the use of
3015 this feature is in deciding when to break out of a \c{%rep}
3016 preprocessor loop: see \k{rep} for a detailed example.
3018 The expression given to \c{%if}, and its counterpart \i\c{%elif}, is
3019 a critical expression (see \k{crit}).
3021 \c{%if} extends the normal NASM expression syntax, by providing a
3022 set of \i{relational operators} which are not normally available in
3023 expressions. The operators \i\c{=}, \i\c{<}, \i\c{>}, \i\c{<=},
3024 \i\c{>=} and \i\c{<>} test equality, less-than, greater-than,
3025 less-or-equal, greater-or-equal and not-equal respectively. The
3026 C-like forms \i\c{==} and \i\c{!=} are supported as alternative
3027 forms of \c{=} and \c{<>}. In addition, low-priority logical
3028 operators \i\c{&&}, \i\c{^^} and \i\c{||} are provided, supplying
3029 \i{logical AND}, \i{logical XOR} and \i{logical OR}. These work like
3030 the C logical operators (although C has no logical XOR), in that
3031 they always return either 0 or 1, and treat any non-zero input as 1
3032 (so that \c{^^}, for example, returns 1 if exactly one of its inputs
3033 is zero, and 0 otherwise). The relational operators also return 1
3034 for true and 0 for false.
3036 Like other \c{%if} constructs, \c{%if} has a counterpart
3037 \i\c{%elif}, and negative forms \i\c{%ifn} and \i\c{%elifn}.
3039 \S{ifidn} \i\c{%ifidn} and \i\c{%ifidni}: Testing Exact Text
3040 Identity\I{testing, exact text identity}
3042 The construct \c{%ifidn text1,text2} will cause the subsequent code
3043 to be assembled if and only if \c{text1} and \c{text2}, after
3044 expanding single-line macros, are identical pieces of text.
3045 Differences in white space are not counted.
3047 \c{%ifidni} is similar to \c{%ifidn}, but is \i{case-insensitive}.
3049 For example, the following macro pushes a register or number on the
3050 stack, and allows you to treat \c{IP} as a real register:
3052 \c %macro pushparam 1
3063 Like other \c{%if} constructs, \c{%ifidn} has a counterpart
3064 \i\c{%elifidn}, and negative forms \i\c{%ifnidn} and \i\c{%elifnidn}.
3065 Similarly, \c{%ifidni} has counterparts \i\c{%elifidni},
3066 \i\c{%ifnidni} and \i\c{%elifnidni}.
3068 \S{iftyp} \i\c{%ifid}, \i\c{%ifnum}, \i\c{%ifstr}: Testing Token
3069 Types\I{testing, token types}
3071 Some macros will want to perform different tasks depending on
3072 whether they are passed a number, a string, or an identifier. For
3073 example, a string output macro might want to be able to cope with
3074 being passed either a string constant or a pointer to an existing
3077 The conditional assembly construct \c{%ifid}, taking one parameter
3078 (which may be blank), assembles the subsequent code if and only if
3079 the first token in the parameter exists and is an identifier.
3080 \c{%ifnum} works similarly, but tests for the token being a numeric
3081 constant; \c{%ifstr} tests for it being a string.
3083 For example, the \c{writefile} macro defined in \k{mlmacgre} can be
3084 extended to take advantage of \c{%ifstr} in the following fashion:
3086 \c %macro writefile 2-3+
3095 \c %%endstr: mov dx,%%str
3096 \c mov cx,%%endstr-%%str
3107 Then the \c{writefile} macro can cope with being called in either of
3108 the following two ways:
3110 \c writefile [file], strpointer, length
3111 \c writefile [file], "hello", 13, 10
3113 In the first, \c{strpointer} is used as the address of an
3114 already-declared string, and \c{length} is used as its length; in
3115 the second, a string is given to the macro, which therefore declares
3116 it itself and works out the address and length for itself.
3118 Note the use of \c{%if} inside the \c{%ifstr}: this is to detect
3119 whether the macro was passed two arguments (so the string would be a
3120 single string constant, and \c{db %2} would be adequate) or more (in
3121 which case, all but the first two would be lumped together into
3122 \c{%3}, and \c{db %2,%3} would be required).
3124 The usual \I\c{%elifid}\I\c{%elifnum}\I\c{%elifstr}\c{%elif}...,
3125 \I\c{%ifnid}\I\c{%ifnnum}\I\c{%ifnstr}\c{%ifn}..., and
3126 \I\c{%elifnid}\I\c{%elifnnum}\I\c{%elifnstr}\c{%elifn}... versions
3127 exist for each of \c{%ifid}, \c{%ifnum} and \c{%ifstr}.
3129 \S{iftoken} \i\c{%iftoken}: Test for a Single Token
3131 Some macros will want to do different things depending on if it is
3132 passed a single token (e.g. paste it to something else using \c{%+})
3133 versus a multi-token sequence.
3135 The conditional assembly construct \c{%iftoken} assembles the
3136 subsequent code if and only if the expanded parameters consist of
3137 exactly one token, possibly surrounded by whitespace.
3143 will assemble the subsequent code, but
3147 will not, since \c{-1} contains two tokens: the unary minus operator
3148 \c{-}, and the number \c{1}.
3150 The usual \i\c{%eliftoken}, \i\c\{%ifntoken}, and \i\c{%elifntoken}
3151 variants are also provided.
3153 \S{ifempty} \i\c{%ifempty}: Test for Empty Expansion
3155 The conditional assembly construct \c{%ifempty} assembles the
3156 subsequent code if and only if the expanded parameters do not contain
3157 any tokens at all, whitespace excepted.
3159 The usual \i\c{%elifempty}, \i\c\{%ifnempty}, and \i\c{%elifnempty}
3160 variants are also provided.
3162 \S{ifenv} \i\c{%ifenv}: Test If Environment Variable Exists
3164 The conditional assembly construct \c{%ifenv} assembles the
3165 subsequent code if and only if the environment variable referenced by
3166 the \c{%!<env>} directive exists.
3168 The usual \i\c{%elifenv}, \i\c\{%ifnenv}, and \i\c{%elifnenv}
3169 variants are also provided.
3171 Just as for \c{%!<env>} the argument should be written as a string if
3172 it contains characters that would not be legal in an identifier. See
3175 \H{rep} \i{Preprocessor Loops}\I{repeating code}: \i\c{%rep}
3177 NASM's \c{TIMES} prefix, though useful, cannot be used to invoke a
3178 multi-line macro multiple times, because it is processed by NASM
3179 after macros have already been expanded. Therefore NASM provides
3180 another form of loop, this time at the preprocessor level: \c{%rep}.
3182 The directives \c{%rep} and \i\c{%endrep} (\c{%rep} takes a numeric
3183 argument, which can be an expression; \c{%endrep} takes no
3184 arguments) can be used to enclose a chunk of code, which is then
3185 replicated as many times as specified by the preprocessor:
3189 \c inc word [table+2*i]
3193 This will generate a sequence of 64 \c{INC} instructions,
3194 incrementing every word of memory from \c{[table]} to
3197 For more complex termination conditions, or to break out of a repeat
3198 loop part way along, you can use the \i\c{%exitrep} directive to
3199 terminate the loop, like this:
3214 \c fib_number equ ($-fibonacci)/2
3216 This produces a list of all the Fibonacci numbers that will fit in
3217 16 bits. Note that a maximum repeat count must still be given to
3218 \c{%rep}. This is to prevent the possibility of NASM getting into an
3219 infinite loop in the preprocessor, which (on multitasking or
3220 multi-user systems) would typically cause all the system memory to
3221 be gradually used up and other applications to start crashing.
3223 Note a maximum repeat count is limited by 62 bit number, though it
3224 is hardly possible that you ever need anything bigger.
3227 \H{while} \i{Conditional Loops}: \i\c{%while}
3229 The directives \c{%while} and \i\c{%endwhile} combine preprocessor
3230 loops with conditional assembly, allowing the enclosed chunk of
3231 code to be replicated as long as certain conditions are met:
3233 \c %while<condition>
3234 \c ; some code which only repeats while <condition> is met
3237 \S{exitwhile} Exiting Conditional Loops: \i\c{%exitwhile}
3239 Conditional loops can be arbitrarily terminated with the
3240 \i\c{%exitwhile} directive.
3244 \c %while<condition>
3245 \c %if<some other condition>
3248 \c ; some code which only repeats while <condition> is met
3252 \H{files} Source Files and Dependencies
3254 These commands allow you to split your sources into multiple files.
3256 \S{include} \i\c{%include}: \i{Including Other Files}
3258 Using, once again, a very similar syntax to the C preprocessor,
3259 NASM's preprocessor lets you include other source files into your
3260 code. This is done by the use of the \i\c{%include} directive:
3262 \c %include "macros.mac"
3264 will include the contents of the file \c{macros.mac} into the source
3265 file containing the \c{%include} directive.
3267 Include files are \I{searching for include files}searched for in the
3268 current directory (the directory you're in when you run NASM, as
3269 opposed to the location of the NASM executable or the location of
3270 the source file), plus any directories specified on the NASM command
3271 line using the \c{-i} option.
3273 The standard C idiom for preventing a file being included more than
3274 once is just as applicable in NASM: if the file \c{macros.mac} has
3277 \c %ifndef MACROS_MAC
3278 \c %define MACROS_MAC
3279 \c ; now define some macros
3282 then including the file more than once will not cause errors,
3283 because the second time the file is included nothing will happen
3284 because the macro \c{MACROS_MAC} will already be defined.
3286 You can force a file to be included even if there is no \c{%include}
3287 directive that explicitly includes it, by using the \i\c{-p} option
3288 on the NASM command line (see \k{opt-p}).
3291 \S{pathsearch} \i\c{%pathsearch}: Search the Include Path
3293 The \c{%pathsearch} directive takes a single-line macro name and a
3294 filename, and declare or redefines the specified single-line macro to
3295 be the include-path-resolved version of the filename, if the file
3296 exists (otherwise, it is passed unchanged.)
3300 \c %pathsearch MyFoo "foo.bin"
3302 ... with \c{-Ibins/} in the include path may end up defining the macro
3303 \c{MyFoo} to be \c{"bins/foo.bin"}.
3306 \S{depend} \i\c{%depend}: Add Dependent Files
3308 The \c{%depend} directive takes a filename and adds it to the list of
3309 files to be emitted as dependency generation when the \c{-M} options
3310 and its relatives (see \k{opt-M}) are used. It produces no output.
3312 This is generally used in conjunction with \c{%pathsearch}. For
3313 example, a simplified version of the standard macro wrapper for the
3314 \c{INCBIN} directive looks like:
3316 \c %imacro incbin 1-2+ 0
3317 \c %pathsearch dep %1
3322 This first resolves the location of the file into the macro \c{dep},
3323 then adds it to the dependency lists, and finally issues the
3324 assembler-level \c{INCBIN} directive.
3327 \S{use} \i\c{%use}: Include Standard Macro Package
3329 The \c{%use} directive is similar to \c{%include}, but rather than
3330 including the contents of a file, it includes a named standard macro
3331 package. The standard macro packages are part of NASM, and are
3332 described in \k{macropkg}.
3334 Unlike the \c{%include} directive, package names for the \c{%use}
3335 directive do not require quotes, but quotes are permitted. In NASM
3336 2.04 and 2.05 the unquoted form would be macro-expanded; this is no
3337 longer true. Thus, the following lines are equivalent:
3342 Standard macro packages are protected from multiple inclusion. When a
3343 standard macro package is used, a testable single-line macro of the
3344 form \c{__USE_}\e{package}\c{__} is also defined, see \k{use_def}.
3346 \H{ctxstack} The \i{Context Stack}
3348 Having labels that are local to a macro definition is sometimes not
3349 quite powerful enough: sometimes you want to be able to share labels
3350 between several macro calls. An example might be a \c{REPEAT} ...
3351 \c{UNTIL} loop, in which the expansion of the \c{REPEAT} macro
3352 would need to be able to refer to a label which the \c{UNTIL} macro
3353 had defined. However, for such a macro you would also want to be
3354 able to nest these loops.
3356 NASM provides this level of power by means of a \e{context stack}.
3357 The preprocessor maintains a stack of \e{contexts}, each of which is
3358 characterized by a name. You add a new context to the stack using
3359 the \i\c{%push} directive, and remove one using \i\c{%pop}. You can
3360 define labels that are local to a particular context on the stack.
3363 \S{pushpop} \i\c{%push} and \i\c{%pop}: \I{creating
3364 contexts}\I{removing contexts}Creating and Removing Contexts
3366 The \c{%push} directive is used to create a new context and place it
3367 on the top of the context stack. \c{%push} takes an optional argument,
3368 which is the name of the context. For example:
3372 This pushes a new context called \c{foobar} on the stack. You can have
3373 several contexts on the stack with the same name: they can still be
3374 distinguished. If no name is given, the context is unnamed (this is
3375 normally used when both the \c{%push} and the \c{%pop} are inside a
3376 single macro definition.)
3378 The directive \c{%pop}, taking one optional argument, removes the top
3379 context from the context stack and destroys it, along with any
3380 labels associated with it. If an argument is given, it must match the
3381 name of the current context, otherwise it will issue an error.
3384 \S{ctxlocal} \i{Context-Local Labels}
3386 Just as the usage \c{%%foo} defines a label which is local to the
3387 particular macro call in which it is used, the usage \I{%$}\c{%$foo}
3388 is used to define a label which is local to the context on the top
3389 of the context stack. So the \c{REPEAT} and \c{UNTIL} example given
3390 above could be implemented by means of:
3406 and invoked by means of, for example,
3414 which would scan every fourth byte of a string in search of the byte
3417 If you need to define, or access, labels local to the context
3418 \e{below} the top one on the stack, you can use \I{%$$}\c{%$$foo}, or
3419 \c{%$$$foo} for the context below that, and so on.
3422 \S{ctxdefine} \i{Context-Local Single-Line Macros}
3424 NASM also allows you to define single-line macros which are local to
3425 a particular context, in just the same way:
3427 \c %define %$localmac 3
3429 will define the single-line macro \c{%$localmac} to be local to the
3430 top context on the stack. Of course, after a subsequent \c{%push},
3431 it can then still be accessed by the name \c{%$$localmac}.
3434 \S{ctxfallthrough} \i{Context Fall-Through Lookup}
3436 Context fall-through lookup (automatic searching of outer contexts)
3437 is a feature that was added in NASM version 0.98.03. Unfortunately,
3438 this feature is unintuitive and can result in buggy code that would
3439 have otherwise been prevented by NASM's error reporting. As a result,
3440 this feature has been \e{deprecated}. NASM version 2.09 will issue a
3441 warning when usage of this \e{deprecated} feature is detected. Starting
3442 with NASM version 2.10, usage of this \e{deprecated} feature will simply
3443 result in an \e{expression syntax error}.
3445 An example usage of this \e{deprecated} feature follows:
3449 \c %assign %$external 1
3451 \c %assign %$internal 1
3452 \c mov eax, %$external
3453 \c mov eax, %$internal
3458 As demonstrated, \c{%$external} is being defined in the \c{ctx1}
3459 context and referenced within the \c{ctx2} context. With context
3460 fall-through lookup, referencing an undefined context-local macro
3461 like this implicitly searches through all outer contexts until a match
3462 is made or isn't found in any context. As a result, \c{%$external}
3463 referenced within the \c{ctx2} context would implicitly use \c{%$external}
3464 as defined in \c{ctx1}. Most people would expect NASM to issue an error in
3465 this situation because \c{%$external} was never defined within \c{ctx2} and also
3466 isn't qualified with the proper context depth, \c{%$$external}.
3468 Here is a revision of the above example with proper context depth:
3472 \c %assign %$external 1
3474 \c %assign %$internal 1
3475 \c mov eax, %$$external
3476 \c mov eax, %$internal
3481 As demonstrated, \c{%$external} is still being defined in the \c{ctx1}
3482 context and referenced within the \c{ctx2} context. However, the
3483 reference to \c{%$external} within \c{ctx2} has been fully qualified with
3484 the proper context depth, \c{%$$external}, and thus is no longer ambiguous,
3485 unintuitive or erroneous.
3488 \S{ctxrepl} \i\c{%repl}: \I{renaming contexts}Renaming a Context
3490 If you need to change the name of the top context on the stack (in
3491 order, for example, to have it respond differently to \c{%ifctx}),
3492 you can execute a \c{%pop} followed by a \c{%push}; but this will
3493 have the side effect of destroying all context-local labels and
3494 macros associated with the context that was just popped.
3496 NASM provides the directive \c{%repl}, which \e{replaces} a context
3497 with a different name, without touching the associated macros and
3498 labels. So you could replace the destructive code
3503 with the non-destructive version \c{%repl newname}.
3506 \S{blockif} Example Use of the \i{Context Stack}: \i{Block IFs}
3508 This example makes use of almost all the context-stack features,
3509 including the conditional-assembly construct \i\c{%ifctx}, to
3510 implement a block IF statement as a set of macros.
3526 \c %error "expected `if' before `else'"
3540 \c %error "expected `if' or `else' before `endif'"
3545 This code is more robust than the \c{REPEAT} and \c{UNTIL} macros
3546 given in \k{ctxlocal}, because it uses conditional assembly to check
3547 that the macros are issued in the right order (for example, not
3548 calling \c{endif} before \c{if}) and issues a \c{%error} if they're
3551 In addition, the \c{endif} macro has to be able to cope with the two
3552 distinct cases of either directly following an \c{if}, or following
3553 an \c{else}. It achieves this, again, by using conditional assembly
3554 to do different things depending on whether the context on top of
3555 the stack is \c{if} or \c{else}.
3557 The \c{else} macro has to preserve the context on the stack, in
3558 order to have the \c{%$ifnot} referred to by the \c{if} macro be the
3559 same as the one defined by the \c{endif} macro, but has to change
3560 the context's name so that \c{endif} will know there was an
3561 intervening \c{else}. It does this by the use of \c{%repl}.
3563 A sample usage of these macros might look like:
3585 The block-\c{IF} macros handle nesting quite happily, by means of
3586 pushing another context, describing the inner \c{if}, on top of the
3587 one describing the outer \c{if}; thus \c{else} and \c{endif} always
3588 refer to the last unmatched \c{if} or \c{else}.
3591 \H{stackrel} \i{Stack Relative Preprocessor Directives}
3593 The following preprocessor directives provide a way to use
3594 labels to refer to local variables allocated on the stack.
3596 \b\c{%arg} (see \k{arg})
3598 \b\c{%stacksize} (see \k{stacksize})
3600 \b\c{%local} (see \k{local})
3603 \S{arg} \i\c{%arg} Directive
3605 The \c{%arg} directive is used to simplify the handling of
3606 parameters passed on the stack. Stack based parameter passing
3607 is used by many high level languages, including C, C++ and Pascal.
3609 While NASM has macros which attempt to duplicate this
3610 functionality (see \k{16cmacro}), the syntax is not particularly
3611 convenient to use and is not TASM compatible. Here is an example
3612 which shows the use of \c{%arg} without any external macros:
3616 \c %push mycontext ; save the current context
3617 \c %stacksize large ; tell NASM to use bp
3618 \c %arg i:word, j_ptr:word
3625 \c %pop ; restore original context
3627 This is similar to the procedure defined in \k{16cmacro} and adds
3628 the value in i to the value pointed to by j_ptr and returns the
3629 sum in the ax register. See \k{pushpop} for an explanation of
3630 \c{push} and \c{pop} and the use of context stacks.
3633 \S{stacksize} \i\c{%stacksize} Directive
3635 The \c{%stacksize} directive is used in conjunction with the
3636 \c{%arg} (see \k{arg}) and the \c{%local} (see \k{local}) directives.
3637 It tells NASM the default size to use for subsequent \c{%arg} and
3638 \c{%local} directives. The \c{%stacksize} directive takes one
3639 required argument which is one of \c{flat}, \c{flat64}, \c{large} or \c{small}.
3643 This form causes NASM to use stack-based parameter addressing
3644 relative to \c{ebp} and it assumes that a near form of call was used
3645 to get to this label (i.e. that \c{eip} is on the stack).
3647 \c %stacksize flat64
3649 This form causes NASM to use stack-based parameter addressing
3650 relative to \c{rbp} and it assumes that a near form of call was used
3651 to get to this label (i.e. that \c{rip} is on the stack).
3655 This form uses \c{bp} to do stack-based parameter addressing and
3656 assumes that a far form of call was used to get to this address
3657 (i.e. that \c{ip} and \c{cs} are on the stack).
3661 This form also uses \c{bp} to address stack parameters, but it is
3662 different from \c{large} because it also assumes that the old value
3663 of bp is pushed onto the stack (i.e. it expects an \c{ENTER}
3664 instruction). In other words, it expects that \c{bp}, \c{ip} and
3665 \c{cs} are on the top of the stack, underneath any local space which
3666 may have been allocated by \c{ENTER}. This form is probably most
3667 useful when used in combination with the \c{%local} directive
3671 \S{local} \i\c{%local} Directive
3673 The \c{%local} directive is used to simplify the use of local
3674 temporary stack variables allocated in a stack frame. Automatic
3675 local variables in C are an example of this kind of variable. The
3676 \c{%local} directive is most useful when used with the \c{%stacksize}
3677 (see \k{stacksize} and is also compatible with the \c{%arg} directive
3678 (see \k{arg}). It allows simplified reference to variables on the
3679 stack which have been allocated typically by using the \c{ENTER}
3681 \# (see \k{insENTER} for a description of that instruction).
3682 An example of its use is the following:
3686 \c %push mycontext ; save the current context
3687 \c %stacksize small ; tell NASM to use bp
3688 \c %assign %$localsize 0 ; see text for explanation
3689 \c %local old_ax:word, old_dx:word
3691 \c enter %$localsize,0 ; see text for explanation
3692 \c mov [old_ax],ax ; swap ax & bx
3693 \c mov [old_dx],dx ; and swap dx & cx
3698 \c leave ; restore old bp
3701 \c %pop ; restore original context
3703 The \c{%$localsize} variable is used internally by the
3704 \c{%local} directive and \e{must} be defined within the
3705 current context before the \c{%local} directive may be used.
3706 Failure to do so will result in one expression syntax error for
3707 each \c{%local} variable declared. It then may be used in
3708 the construction of an appropriately sized ENTER instruction
3709 as shown in the example.
3712 \H{pperror} Reporting \i{User-Defined Errors}: \i\c{%error}, \i\c{%warning}, \i\c{%fatal}
3714 The preprocessor directive \c{%error} will cause NASM to report an
3715 error if it occurs in assembled code. So if other users are going to
3716 try to assemble your source files, you can ensure that they define the
3717 right macros by means of code like this:
3722 \c ; do some different setup
3724 \c %error "Neither F1 nor F2 was defined."
3727 Then any user who fails to understand the way your code is supposed
3728 to be assembled will be quickly warned of their mistake, rather than
3729 having to wait until the program crashes on being run and then not
3730 knowing what went wrong.
3732 Similarly, \c{%warning} issues a warning, but allows assembly to continue:
3737 \c ; do some different setup
3739 \c %warning "Neither F1 nor F2 was defined, assuming F1."
3743 \c{%error} and \c{%warning} are issued only on the final assembly
3744 pass. This makes them safe to use in conjunction with tests that
3745 depend on symbol values.
3747 \c{%fatal} terminates assembly immediately, regardless of pass. This
3748 is useful when there is no point in continuing the assembly further,
3749 and doing so is likely just going to cause a spew of confusing error
3752 It is optional for the message string after \c{%error}, \c{%warning}
3753 or \c{%fatal} to be quoted. If it is \e{not}, then single-line macros
3754 are expanded in it, which can be used to display more information to
3755 the user. For example:
3758 \c %assign foo_over foo-64
3759 \c %error foo is foo_over bytes too large
3763 \H{otherpreproc} \i{Other Preprocessor Directives}
3765 NASM also has preprocessor directives which allow access to
3766 information from external sources. Currently they include:
3768 \b\c{%line} enables NASM to correctly handle the output of another
3769 preprocessor (see \k{line}).
3771 \b\c{%!} enables NASM to read in the value of an environment variable,
3772 which can then be used in your program (see \k{getenv}).
3774 \S{line} \i\c{%line} Directive
3776 The \c{%line} directive is used to notify NASM that the input line
3777 corresponds to a specific line number in another file. Typically
3778 this other file would be an original source file, with the current
3779 NASM input being the output of a pre-processor. The \c{%line}
3780 directive allows NASM to output messages which indicate the line
3781 number of the original source file, instead of the file that is being
3784 This preprocessor directive is not generally of use to programmers,
3785 by may be of interest to preprocessor authors. The usage of the
3786 \c{%line} preprocessor directive is as follows:
3788 \c %line nnn[+mmm] [filename]
3790 In this directive, \c{nnn} identifies the line of the original source
3791 file which this line corresponds to. \c{mmm} is an optional parameter
3792 which specifies a line increment value; each line of the input file
3793 read in is considered to correspond to \c{mmm} lines of the original
3794 source file. Finally, \c{filename} is an optional parameter which
3795 specifies the file name of the original source file.
3797 After reading a \c{%line} preprocessor directive, NASM will report
3798 all file name and line numbers relative to the values specified
3802 \S{getenv} \i\c{%!}\c{<env>}: Read an environment variable.
3804 The \c{%!<env>} directive makes it possible to read the value of an
3805 environment variable at assembly time. This could, for example, be used
3806 to store the contents of an environment variable into a string, which
3807 could be used at some other point in your code.
3809 For example, suppose that you have an environment variable \c{FOO}, and
3810 you want the contents of \c{FOO} to be embedded in your program. You
3811 could do that as follows:
3813 \c %defstr FOO %!FOO
3815 See \k{defstr} for notes on the \c{%defstr} directive.
3817 If the name of the environment variable contains non-identifier
3818 characters, you can use string quotes to surround the name of the
3819 variable, for example:
3821 \c %defstr C_colon %!'C:'
3824 \S{final} \i\c{%final} Directive
3826 The \c{%final} directive is used to delay preprocessing of a line
3827 until all other "normal" preprocessing is complete. Multiple
3828 \c{%final} directives are processed in the opposite order of their
3829 declaration, last one first and first one last.
3832 \H{comment} Comment Blocks: \i\c{%comment}
3834 The \c{%comment} and \c{%endcomment} directives are used to specify
3835 a block of commented (i.e. unprocessed) code/text. Everything between
3836 \c{%comment} and \c{%endcomment} will be ignored by the preprocessor.
3839 \c ; some code, text or data to be ignored
3843 \H{stdmac} \i{Standard Macros}
3845 NASM defines a set of standard macros, which are already defined
3846 when it starts to process any source file. If you really need a
3847 program to be assembled with no pre-defined macros, you can use the
3848 \i\c{%clear} directive to empty the preprocessor of everything but
3849 context-local preprocessor variables and single-line macros.
3851 Most \i{user-level assembler directives} (see \k{directive}) are
3852 implemented as macros which invoke primitive directives; these are
3853 described in \k{directive}. The rest of the standard macro set is
3857 \S{stdmacver} \i{NASM Version} Macros
3859 The single-line macros \i\c{__NASM_MAJOR__}, \i\c{__NASM_MINOR__},
3860 \i\c{__NASM_SUBMINOR__} and \i\c{___NASM_PATCHLEVEL__} expand to the
3861 major, minor, subminor and patch level parts of the \i{version
3862 number of NASM} being used. So, under NASM 0.98.32p1 for
3863 example, \c{__NASM_MAJOR__} would be defined to be 0, \c{__NASM_MINOR__}
3864 would be defined as 98, \c{__NASM_SUBMINOR__} would be defined to 32,
3865 and \c{___NASM_PATCHLEVEL__} would be defined as 1.
3867 Additionally, the macro \i\c{__NASM_SNAPSHOT__} is defined for
3868 automatically generated snapshot releases \e{only}.
3871 \S{stdmacverid} \i\c{__NASM_VERSION_ID__}: \i{NASM Version ID}
3873 The single-line macro \c{__NASM_VERSION_ID__} expands to a dword integer
3874 representing the full version number of the version of nasm being used.
3875 The value is the equivalent to \c{__NASM_MAJOR__}, \c{__NASM_MINOR__},
3876 \c{__NASM_SUBMINOR__} and \c{___NASM_PATCHLEVEL__} concatenated to
3877 produce a single doubleword. Hence, for 0.98.32p1, the returned number
3878 would be equivalent to:
3886 Note that the above lines are generate exactly the same code, the second
3887 line is used just to give an indication of the order that the separate
3888 values will be present in memory.
3891 \S{stdmacverstr} \i\c{__NASM_VER__}: \i{NASM Version string}
3893 The single-line macro \c{__NASM_VER__} expands to a string which defines
3894 the version number of nasm being used. So, under NASM 0.98.32 for example,
3903 \S{fileline} \i\c{__FILE__} and \i\c{__LINE__}: File Name and Line Number
3905 Like the C preprocessor, NASM allows the user to find out the file
3906 name and line number containing the current instruction. The macro
3907 \c{__FILE__} expands to a string constant giving the name of the
3908 current input file (which may change through the course of assembly
3909 if \c{%include} directives are used), and \c{__LINE__} expands to a
3910 numeric constant giving the current line number in the input file.
3912 These macros could be used, for example, to communicate debugging
3913 information to a macro, since invoking \c{__LINE__} inside a macro
3914 definition (either single-line or multi-line) will return the line
3915 number of the macro \e{call}, rather than \e{definition}. So to
3916 determine where in a piece of code a crash is occurring, for
3917 example, one could write a routine \c{stillhere}, which is passed a
3918 line number in \c{EAX} and outputs something like `line 155: still
3919 here'. You could then write a macro
3921 \c %macro notdeadyet 0
3930 and then pepper your code with calls to \c{notdeadyet} until you
3931 find the crash point.
3934 \S{bitsm} \i\c{__BITS__}: Current BITS Mode
3936 The \c{__BITS__} standard macro is updated every time that the BITS mode is
3937 set using the \c{BITS XX} or \c{[BITS XX]} directive, where XX is a valid mode
3938 number of 16, 32 or 64. \c{__BITS__} receives the specified mode number and
3939 makes it globally available. This can be very useful for those who utilize
3940 mode-dependent macros.
3942 \S{ofmtm} \i\c{__OUTPUT_FORMAT__}: Current Output Format
3944 The \c{__OUTPUT_FORMAT__} standard macro holds the current Output Format,
3945 as given by the \c{-f} option or NASM's default. Type \c{nasm -hf} for a
3948 \c %ifidn __OUTPUT_FORMAT__, win32
3949 \c %define NEWLINE 13, 10
3950 \c %elifidn __OUTPUT_FORMAT__, elf32
3951 \c %define NEWLINE 10
3955 \S{datetime} Assembly Date and Time Macros
3957 NASM provides a variety of macros that represent the timestamp of the
3960 \b The \i\c{__DATE__} and \i\c{__TIME__} macros give the assembly date and
3961 time as strings, in ISO 8601 format (\c{"YYYY-MM-DD"} and \c{"HH:MM:SS"},
3964 \b The \i\c{__DATE_NUM__} and \i\c{__TIME_NUM__} macros give the assembly
3965 date and time in numeric form; in the format \c{YYYYMMDD} and
3966 \c{HHMMSS} respectively.
3968 \b The \i\c{__UTC_DATE__} and \i\c{__UTC_TIME__} macros give the assembly
3969 date and time in universal time (UTC) as strings, in ISO 8601 format
3970 (\c{"YYYY-MM-DD"} and \c{"HH:MM:SS"}, respectively.) If the host
3971 platform doesn't provide UTC time, these macros are undefined.
3973 \b The \i\c{__UTC_DATE_NUM__} and \i\c{__UTC_TIME_NUM__} macros give the
3974 assembly date and time universal time (UTC) in numeric form; in the
3975 format \c{YYYYMMDD} and \c{HHMMSS} respectively. If the
3976 host platform doesn't provide UTC time, these macros are
3979 \b The \c{__POSIX_TIME__} macro is defined as a number containing the
3980 number of seconds since the POSIX epoch, 1 January 1970 00:00:00 UTC;
3981 excluding any leap seconds. This is computed using UTC time if
3982 available on the host platform, otherwise it is computed using the
3983 local time as if it was UTC.
3985 All instances of time and date macros in the same assembly session
3986 produce consistent output. For example, in an assembly session
3987 started at 42 seconds after midnight on January 1, 2010 in Moscow
3988 (timezone UTC+3) these macros would have the following values,
3989 assuming, of course, a properly configured environment with a correct
3992 \c __DATE__ "2010-01-01"
3993 \c __TIME__ "00:00:42"
3994 \c __DATE_NUM__ 20100101
3995 \c __TIME_NUM__ 000042
3996 \c __UTC_DATE__ "2009-12-31"
3997 \c __UTC_TIME__ "21:00:42"
3998 \c __UTC_DATE_NUM__ 20091231
3999 \c __UTC_TIME_NUM__ 210042
4000 \c __POSIX_TIME__ 1262293242
4003 \S{use_def} \I\c{__USE_*__}\c{__USE_}\e{package}\c{__}: Package
4006 When a standard macro package (see \k{macropkg}) is included with the
4007 \c{%use} directive (see \k{use}), a single-line macro of the form
4008 \c{__USE_}\e{package}\c{__} is automatically defined. This allows
4009 testing if a particular package is invoked or not.
4011 For example, if the \c{altreg} package is included (see
4012 \k{pkg_altreg}), then the macro \c{__USE_ALTREG__} is defined.
4015 \S{pass_macro} \i\c{__PASS__}: Assembly Pass
4017 The macro \c{__PASS__} is defined to be \c{1} on preparatory passes,
4018 and \c{2} on the final pass. In preprocess-only mode, it is set to
4019 \c{3}, and when running only to generate dependencies (due to the
4020 \c{-M} or \c{-MG} option, see \k{opt-M}) it is set to \c{0}.
4022 \e{Avoid using this macro if at all possible. It is tremendously easy
4023 to generate very strange errors by misusing it, and the semantics may
4024 change in future versions of NASM.}
4027 \S{struc} \i\c{STRUC} and \i\c{ENDSTRUC}: \i{Declaring Structure} Data Types
4029 The core of NASM contains no intrinsic means of defining data
4030 structures; instead, the preprocessor is sufficiently powerful that
4031 data structures can be implemented as a set of macros. The macros
4032 \c{STRUC} and \c{ENDSTRUC} are used to define a structure data type.
4034 \c{STRUC} takes one or two parameters. The first parameter is the name
4035 of the data type. The second, optional parameter is the base offset of
4036 the structure. The name of the data type is defined as a symbol with
4037 the value of the base offset, and the name of the data type with the
4038 suffix \c{_size} appended to it is defined as an \c{EQU} giving the
4039 size of the structure. Once \c{STRUC} has been issued, you are
4040 defining the structure, and should define fields using the \c{RESB}
4041 family of pseudo-instructions, and then invoke \c{ENDSTRUC} to finish
4044 For example, to define a structure called \c{mytype} containing a
4045 longword, a word, a byte and a string of bytes, you might code
4056 The above code defines six symbols: \c{mt_long} as 0 (the offset
4057 from the beginning of a \c{mytype} structure to the longword field),
4058 \c{mt_word} as 4, \c{mt_byte} as 6, \c{mt_str} as 7, \c{mytype_size}
4059 as 39, and \c{mytype} itself as zero.
4061 The reason why the structure type name is defined at zero by default
4062 is a side effect of allowing structures to work with the local label
4063 mechanism: if your structure members tend to have the same names in
4064 more than one structure, you can define the above structure like this:
4075 This defines the offsets to the structure fields as \c{mytype.long},
4076 \c{mytype.word}, \c{mytype.byte} and \c{mytype.str}.
4078 NASM, since it has no \e{intrinsic} structure support, does not
4079 support any form of period notation to refer to the elements of a
4080 structure once you have one (except the above local-label notation),
4081 so code such as \c{mov ax,[mystruc.mt_word]} is not valid.
4082 \c{mt_word} is a constant just like any other constant, so the
4083 correct syntax is \c{mov ax,[mystruc+mt_word]} or \c{mov
4084 ax,[mystruc+mytype.word]}.
4086 Sometimes you only have the address of the structure displaced by an
4087 offset. For example, consider this standard stack frame setup:
4093 In this case, you could access an element by subtracting the offset:
4095 \c mov [ebp - 40 + mytype.word], ax
4097 However, if you do not want to repeat this offset, you can use -40 as
4100 \c struc mytype, -40
4102 And access an element this way:
4104 \c mov [ebp + mytype.word], ax
4107 \S{istruc} \i\c{ISTRUC}, \i\c{AT} and \i\c{IEND}: Declaring
4108 \i{Instances of Structures}
4110 Having defined a structure type, the next thing you typically want
4111 to do is to declare instances of that structure in your data
4112 segment. NASM provides an easy way to do this in the \c{ISTRUC}
4113 mechanism. To declare a structure of type \c{mytype} in a program,
4114 you code something like this:
4119 \c at mt_long, dd 123456
4120 \c at mt_word, dw 1024
4121 \c at mt_byte, db 'x'
4122 \c at mt_str, db 'hello, world', 13, 10, 0
4126 The function of the \c{AT} macro is to make use of the \c{TIMES}
4127 prefix to advance the assembly position to the correct point for the
4128 specified structure field, and then to declare the specified data.
4129 Therefore the structure fields must be declared in the same order as
4130 they were specified in the structure definition.
4132 If the data to go in a structure field requires more than one source
4133 line to specify, the remaining source lines can easily come after
4134 the \c{AT} line. For example:
4136 \c at mt_str, db 123,134,145,156,167,178,189
4139 Depending on personal taste, you can also omit the code part of the
4140 \c{AT} line completely, and start the structure field on the next
4144 \c db 'hello, world'
4148 \S{align} \i\c{ALIGN} and \i\c{ALIGNB}: Data Alignment
4150 The \c{ALIGN} and \c{ALIGNB} macros provides a convenient way to
4151 align code or data on a word, longword, paragraph or other boundary.
4152 (Some assemblers call this directive \i\c{EVEN}.) The syntax of the
4153 \c{ALIGN} and \c{ALIGNB} macros is
4155 \c align 4 ; align on 4-byte boundary
4156 \c align 16 ; align on 16-byte boundary
4157 \c align 8,db 0 ; pad with 0s rather than NOPs
4158 \c align 4,resb 1 ; align to 4 in the BSS
4159 \c alignb 4 ; equivalent to previous line
4161 Both macros require their first argument to be a power of two; they
4162 both compute the number of additional bytes required to bring the
4163 length of the current section up to a multiple of that power of two,
4164 and then apply the \c{TIMES} prefix to their second argument to
4165 perform the alignment.
4167 If the second argument is not specified, the default for \c{ALIGN}
4168 is \c{NOP}, and the default for \c{ALIGNB} is \c{RESB 1}. So if the
4169 second argument is specified, the two macros are equivalent.
4170 Normally, you can just use \c{ALIGN} in code and data sections and
4171 \c{ALIGNB} in BSS sections, and never need the second argument
4172 except for special purposes.
4174 \c{ALIGN} and \c{ALIGNB}, being simple macros, perform no error
4175 checking: they cannot warn you if their first argument fails to be a
4176 power of two, or if their second argument generates more than one
4177 byte of code. In each of these cases they will silently do the wrong
4180 \c{ALIGNB} (or \c{ALIGN} with a second argument of \c{RESB 1}) can
4181 be used within structure definitions:
4198 This will ensure that the structure members are sensibly aligned
4199 relative to the base of the structure.
4201 A final caveat: \c{ALIGN} and \c{ALIGNB} work relative to the
4202 beginning of the \e{section}, not the beginning of the address space
4203 in the final executable. Aligning to a 16-byte boundary when the
4204 section you're in is only guaranteed to be aligned to a 4-byte
4205 boundary, for example, is a waste of effort. Again, NASM does not
4206 check that the section's alignment characteristics are sensible for
4207 the use of \c{ALIGN} or \c{ALIGNB}.
4209 Both \c{ALIGN} and \c{ALIGNB} do call \c{SECTALIGN} macro implicitly.
4210 See \k{sectalign} for details.
4212 See also the \c{smartalign} standard macro package, \k{pkg_smartalign}.
4215 \S{sectalign} \i\c{SECTALIGN}: Section Alignment
4217 The \c{SECTALIGN} macros provides a way to modify alignment attribute
4218 of output file section. Unlike the \c{align=} attribute (which is allowed
4219 at section definition only) the \c{SECTALIGN} macro may be used at any time.
4221 For example the directive
4225 sets the section alignment requirements to 16 bytes. Once increased it can
4226 not be decreased, the magnitude may grow only.
4228 Note that \c{ALIGN} (see \k{align}) calls the \c{SECTALIGN} macro implicitly
4229 so the active section alignment requirements may be updated. This is by default
4230 behaviour, if for some reason you want the \c{ALIGN} do not call \c{SECTALIGN}
4231 at all use the directive
4235 It is still possible to turn in on again by
4240 \C{macropkg} \i{Standard Macro Packages}
4242 The \i\c{%use} directive (see \k{use}) includes one of the standard
4243 macro packages included with the NASM distribution and compiled into
4244 the NASM binary. It operates like the \c{%include} directive (see
4245 \k{include}), but the included contents is provided by NASM itself.
4247 The names of standard macro packages are case insensitive, and can be
4251 \H{pkg_altreg} \i\c{altreg}: \i{Alternate Register Names}
4253 The \c{altreg} standard macro package provides alternate register
4254 names. It provides numeric register names for all registers (not just
4255 \c{R8}-\c{R15}), the Intel-defined aliases \c{R8L}-\c{R15L} for the
4256 low bytes of register (as opposed to the NASM/AMD standard names
4257 \c{R8B}-\c{R15B}), and the names \c{R0H}-\c{R3H} (by analogy with
4258 \c{R0L}-\c{R3L}) for \c{AH}, \c{CH}, \c{DH}, and \c{BH}.
4265 \c mov r0l,r3h ; mov al,bh
4271 \H{pkg_smartalign} \i\c{smartalign}\I{align, smart}: Smart \c{ALIGN} Macro
4273 The \c{smartalign} standard macro package provides for an \i\c{ALIGN}
4274 macro which is more powerful than the default (and
4275 backwards-compatible) one (see \k{align}). When the \c{smartalign}
4276 package is enabled, when \c{ALIGN} is used without a second argument,
4277 NASM will generate a sequence of instructions more efficient than a
4278 series of \c{NOP}. Furthermore, if the padding exceeds a specific
4279 threshold, then NASM will generate a jump over the entire padding
4282 The specific instructions generated can be controlled with the
4283 new \i\c{ALIGNMODE} macro. This macro takes two parameters: one mode,
4284 and an optional jump threshold override. If (for any reason) you need
4285 to turn off the jump completely just set jump threshold value to -1
4286 (or set it to \c{nojmp}). The following modes are possible:
4288 \b \c{generic}: Works on all x86 CPUs and should have reasonable
4289 performance. The default jump threshold is 8. This is the
4292 \b \c{nop}: Pad out with \c{NOP} instructions. The only difference
4293 compared to the standard \c{ALIGN} macro is that NASM can still jump
4294 over a large padding area. The default jump threshold is 16.
4296 \b \c{k7}: Optimize for the AMD K7 (Athlon/Althon XP). These
4297 instructions should still work on all x86 CPUs. The default jump
4300 \b \c{k8}: Optimize for the AMD K8 (Opteron/Althon 64). These
4301 instructions should still work on all x86 CPUs. The default jump
4304 \b \c{p6}: Optimize for Intel CPUs. This uses the long \c{NOP}
4305 instructions first introduced in Pentium Pro. This is incompatible
4306 with all CPUs of family 5 or lower, as well as some VIA CPUs and
4307 several virtualization solutions. The default jump threshold is 16.
4309 The macro \i\c{__ALIGNMODE__} is defined to contain the current
4310 alignment mode. A number of other macros beginning with \c{__ALIGN_}
4311 are used internally by this macro package.
4314 \H{pkg_fp} \i\c\{fp}: Floating-point macros
4316 This packages contains the following floating-point convenience macros:
4318 \c %define Inf __Infinity__
4319 \c %define NaN __QNaN__
4320 \c %define QNaN __QNaN__
4321 \c %define SNaN __SNaN__
4323 \c %define float8(x) __float8__(x)
4324 \c %define float16(x) __float16__(x)
4325 \c %define float32(x) __float32__(x)
4326 \c %define float64(x) __float64__(x)
4327 \c %define float80m(x) __float80m__(x)
4328 \c %define float80e(x) __float80e__(x)
4329 \c %define float128l(x) __float128l__(x)
4330 \c %define float128h(x) __float128h__(x)
4333 \C{directive} \i{Assembler Directives}
4335 NASM, though it attempts to avoid the bureaucracy of assemblers like
4336 MASM and TASM, is nevertheless forced to support a \e{few}
4337 directives. These are described in this chapter.
4339 NASM's directives come in two types: \I{user-level
4340 directives}\e{user-level} directives and \I{primitive
4341 directives}\e{primitive} directives. Typically, each directive has a
4342 user-level form and a primitive form. In almost all cases, we
4343 recommend that users use the user-level forms of the directives,
4344 which are implemented as macros which call the primitive forms.
4346 Primitive directives are enclosed in square brackets; user-level
4349 In addition to the universal directives described in this chapter,
4350 each object file format can optionally supply extra directives in
4351 order to control particular features of that file format. These
4352 \I{format-specific directives}\e{format-specific} directives are
4353 documented along with the formats that implement them, in \k{outfmt}.
4356 \H{bits} \i\c{BITS}: Specifying Target \i{Processor Mode}
4358 The \c{BITS} directive specifies whether NASM should generate code
4359 \I{16-bit mode, versus 32-bit mode}designed to run on a processor
4360 operating in 16-bit mode, 32-bit mode or 64-bit mode. The syntax is
4361 \c{BITS XX}, where XX is 16, 32 or 64.
4363 In most cases, you should not need to use \c{BITS} explicitly. The
4364 \c{aout}, \c{coff}, \c{elf}, \c{macho}, \c{win32} and \c{win64}
4365 object formats, which are designed for use in 32-bit or 64-bit
4366 operating systems, all cause NASM to select 32-bit or 64-bit mode,
4367 respectively, by default. The \c{obj} object format allows you
4368 to specify each segment you define as either \c{USE16} or \c{USE32},
4369 and NASM will set its operating mode accordingly, so the use of the
4370 \c{BITS} directive is once again unnecessary.
4372 The most likely reason for using the \c{BITS} directive is to write
4373 32-bit or 64-bit code in a flat binary file; this is because the \c{bin}
4374 output format defaults to 16-bit mode in anticipation of it being
4375 used most frequently to write DOS \c{.COM} programs, DOS \c{.SYS}
4376 device drivers and boot loader software.
4378 You do \e{not} need to specify \c{BITS 32} merely in order to use
4379 32-bit instructions in a 16-bit DOS program; if you do, the
4380 assembler will generate incorrect code because it will be writing
4381 code targeted at a 32-bit platform, to be run on a 16-bit one.
4383 When NASM is in \c{BITS 16} mode, instructions which use 32-bit
4384 data are prefixed with an 0x66 byte, and those referring to 32-bit
4385 addresses have an 0x67 prefix. In \c{BITS 32} mode, the reverse is
4386 true: 32-bit instructions require no prefixes, whereas instructions
4387 using 16-bit data need an 0x66 and those working on 16-bit addresses
4390 When NASM is in \c{BITS 64} mode, most instructions operate the same
4391 as they do for \c{BITS 32} mode. However, there are 8 more general and
4392 SSE registers, and 16-bit addressing is no longer supported.
4394 The default address size is 64 bits; 32-bit addressing can be selected
4395 with the 0x67 prefix. The default operand size is still 32 bits,
4396 however, and the 0x66 prefix selects 16-bit operand size. The \c{REX}
4397 prefix is used both to select 64-bit operand size, and to access the
4398 new registers. NASM automatically inserts REX prefixes when
4401 When the \c{REX} prefix is used, the processor does not know how to
4402 address the AH, BH, CH or DH (high 8-bit legacy) registers. Instead,
4403 it is possible to access the the low 8-bits of the SP, BP SI and DI
4404 registers as SPL, BPL, SIL and DIL, respectively; but only when the
4407 The \c{BITS} directive has an exactly equivalent primitive form,
4408 \c{[BITS 16]}, \c{[BITS 32]} and \c{[BITS 64]}. The user-level form is
4409 a macro which has no function other than to call the primitive form.
4411 Note that the space is neccessary, e.g. \c{BITS32} will \e{not} work!
4413 \S{USE16 & USE32} \i\c{USE16} & \i\c{USE32}: Aliases for BITS
4415 The `\c{USE16}' and `\c{USE32}' directives can be used in place of
4416 `\c{BITS 16}' and `\c{BITS 32}', for compatibility with other assemblers.
4419 \H{default} \i\c{DEFAULT}: Change the assembler defaults
4421 The \c{DEFAULT} directive changes the assembler defaults. Normally,
4422 NASM defaults to a mode where the programmer is expected to explicitly
4423 specify most features directly. However, this is occationally
4424 obnoxious, as the explicit form is pretty much the only one one wishes
4427 Currently, the only \c{DEFAULT} that is settable is whether or not
4428 registerless instructions in 64-bit mode are \c{RIP}-relative or not.
4429 By default, they are absolute unless overridden with the \i\c{REL}
4430 specifier (see \k{effaddr}). However, if \c{DEFAULT REL} is
4431 specified, \c{REL} is default, unless overridden with the \c{ABS}
4432 specifier, \e{except when used with an FS or GS segment override}.
4434 The special handling of \c{FS} and \c{GS} overrides are due to the
4435 fact that these registers are generally used as thread pointers or
4436 other special functions in 64-bit mode, and generating
4437 \c{RIP}-relative addresses would be extremely confusing.
4439 \c{DEFAULT REL} is disabled with \c{DEFAULT ABS}.
4441 \H{section} \i\c{SECTION} or \i\c{SEGMENT}: Changing and \i{Defining
4444 \I{changing sections}\I{switching between sections}The \c{SECTION}
4445 directive (\c{SEGMENT} is an exactly equivalent synonym) changes
4446 which section of the output file the code you write will be
4447 assembled into. In some object file formats, the number and names of
4448 sections are fixed; in others, the user may make up as many as they
4449 wish. Hence \c{SECTION} may sometimes give an error message, or may
4450 define a new section, if you try to switch to a section that does
4453 The Unix object formats, and the \c{bin} object format (but see
4454 \k{multisec}, all support
4455 the \i{standardized section names} \c{.text}, \c{.data} and \c{.bss}
4456 for the code, data and uninitialized-data sections. The \c{obj}
4457 format, by contrast, does not recognize these section names as being
4458 special, and indeed will strip off the leading period of any section
4462 \S{sectmac} The \i\c{__SECT__} Macro
4464 The \c{SECTION} directive is unusual in that its user-level form
4465 functions differently from its primitive form. The primitive form,
4466 \c{[SECTION xyz]}, simply switches the current target section to the
4467 one given. The user-level form, \c{SECTION xyz}, however, first
4468 defines the single-line macro \c{__SECT__} to be the primitive
4469 \c{[SECTION]} directive which it is about to issue, and then issues
4470 it. So the user-level directive
4474 expands to the two lines
4476 \c %define __SECT__ [SECTION .text]
4479 Users may find it useful to make use of this in their own macros.
4480 For example, the \c{writefile} macro defined in \k{mlmacgre} can be
4481 usefully rewritten in the following more sophisticated form:
4483 \c %macro writefile 2+
4493 \c mov cx,%%endstr-%%str
4500 This form of the macro, once passed a string to output, first
4501 switches temporarily to the data section of the file, using the
4502 primitive form of the \c{SECTION} directive so as not to modify
4503 \c{__SECT__}. It then declares its string in the data section, and
4504 then invokes \c{__SECT__} to switch back to \e{whichever} section
4505 the user was previously working in. It thus avoids the need, in the
4506 previous version of the macro, to include a \c{JMP} instruction to
4507 jump over the data, and also does not fail if, in a complicated
4508 \c{OBJ} format module, the user could potentially be assembling the
4509 code in any of several separate code sections.
4512 \H{absolute} \i\c{ABSOLUTE}: Defining Absolute Labels
4514 The \c{ABSOLUTE} directive can be thought of as an alternative form
4515 of \c{SECTION}: it causes the subsequent code to be directed at no
4516 physical section, but at the hypothetical section starting at the
4517 given absolute address. The only instructions you can use in this
4518 mode are the \c{RESB} family.
4520 \c{ABSOLUTE} is used as follows:
4528 This example describes a section of the PC BIOS data area, at
4529 segment address 0x40: the above code defines \c{kbuf_chr} to be
4530 0x1A, \c{kbuf_free} to be 0x1C, and \c{kbuf} to be 0x1E.
4532 The user-level form of \c{ABSOLUTE}, like that of \c{SECTION},
4533 redefines the \i\c{__SECT__} macro when it is invoked.
4535 \i\c{STRUC} and \i\c{ENDSTRUC} are defined as macros which use
4536 \c{ABSOLUTE} (and also \c{__SECT__}).
4538 \c{ABSOLUTE} doesn't have to take an absolute constant as an
4539 argument: it can take an expression (actually, a \i{critical
4540 expression}: see \k{crit}) and it can be a value in a segment. For
4541 example, a TSR can re-use its setup code as run-time BSS like this:
4543 \c org 100h ; it's a .COM program
4545 \c jmp setup ; setup code comes last
4547 \c ; the resident part of the TSR goes here
4549 \c ; now write the code that installs the TSR here
4553 \c runtimevar1 resw 1
4554 \c runtimevar2 resd 20
4558 This defines some variables `on top of' the setup code, so that
4559 after the setup has finished running, the space it took up can be
4560 re-used as data storage for the running TSR. The symbol `tsr_end'
4561 can be used to calculate the total size of the part of the TSR that
4562 needs to be made resident.
4565 \H{extern} \i\c{EXTERN}: \i{Importing Symbols} from Other Modules
4567 \c{EXTERN} is similar to the MASM directive \c{EXTRN} and the C
4568 keyword \c{extern}: it is used to declare a symbol which is not
4569 defined anywhere in the module being assembled, but is assumed to be
4570 defined in some other module and needs to be referred to by this
4571 one. Not every object-file format can support external variables:
4572 the \c{bin} format cannot.
4574 The \c{EXTERN} directive takes as many arguments as you like. Each
4575 argument is the name of a symbol:
4578 \c extern _sscanf,_fscanf
4580 Some object-file formats provide extra features to the \c{EXTERN}
4581 directive. In all cases, the extra features are used by suffixing a
4582 colon to the symbol name followed by object-format specific text.
4583 For example, the \c{obj} format allows you to declare that the
4584 default segment base of an external should be the group \c{dgroup}
4585 by means of the directive
4587 \c extern _variable:wrt dgroup
4589 The primitive form of \c{EXTERN} differs from the user-level form
4590 only in that it can take only one argument at a time: the support
4591 for multiple arguments is implemented at the preprocessor level.
4593 You can declare the same variable as \c{EXTERN} more than once: NASM
4594 will quietly ignore the second and later redeclarations. You can't
4595 declare a variable as \c{EXTERN} as well as something else, though.
4598 \H{global} \i\c{GLOBAL}: \i{Exporting Symbols} to Other Modules
4600 \c{GLOBAL} is the other end of \c{EXTERN}: if one module declares a
4601 symbol as \c{EXTERN} and refers to it, then in order to prevent
4602 linker errors, some other module must actually \e{define} the
4603 symbol and declare it as \c{GLOBAL}. Some assemblers use the name
4604 \i\c{PUBLIC} for this purpose.
4606 The \c{GLOBAL} directive applying to a symbol must appear \e{before}
4607 the definition of the symbol.
4609 \c{GLOBAL} uses the same syntax as \c{EXTERN}, except that it must
4610 refer to symbols which \e{are} defined in the same module as the
4611 \c{GLOBAL} directive. For example:
4617 \c{GLOBAL}, like \c{EXTERN}, allows object formats to define private
4618 extensions by means of a colon. The \c{elf} object format, for
4619 example, lets you specify whether global data items are functions or
4622 \c global hashlookup:function, hashtable:data
4624 Like \c{EXTERN}, the primitive form of \c{GLOBAL} differs from the
4625 user-level form only in that it can take only one argument at a
4629 \H{common} \i\c{COMMON}: Defining Common Data Areas
4631 The \c{COMMON} directive is used to declare \i\e{common variables}.
4632 A common variable is much like a global variable declared in the
4633 uninitialized data section, so that
4637 is similar in function to
4644 The difference is that if more than one module defines the same
4645 common variable, then at link time those variables will be
4646 \e{merged}, and references to \c{intvar} in all modules will point
4647 at the same piece of memory.
4649 Like \c{GLOBAL} and \c{EXTERN}, \c{COMMON} supports object-format
4650 specific extensions. For example, the \c{obj} format allows common
4651 variables to be NEAR or FAR, and the \c{elf} format allows you to
4652 specify the alignment requirements of a common variable:
4654 \c common commvar 4:near ; works in OBJ
4655 \c common intarray 100:4 ; works in ELF: 4 byte aligned
4657 Once again, like \c{EXTERN} and \c{GLOBAL}, the primitive form of
4658 \c{COMMON} differs from the user-level form only in that it can take
4659 only one argument at a time.
4662 \H{CPU} \i\c{CPU}: Defining CPU Dependencies
4664 The \i\c{CPU} directive restricts assembly to those instructions which
4665 are available on the specified CPU.
4669 \b\c{CPU 8086} Assemble only 8086 instruction set
4671 \b\c{CPU 186} Assemble instructions up to the 80186 instruction set
4673 \b\c{CPU 286} Assemble instructions up to the 286 instruction set
4675 \b\c{CPU 386} Assemble instructions up to the 386 instruction set
4677 \b\c{CPU 486} 486 instruction set
4679 \b\c{CPU 586} Pentium instruction set
4681 \b\c{CPU PENTIUM} Same as 586
4683 \b\c{CPU 686} P6 instruction set
4685 \b\c{CPU PPRO} Same as 686
4687 \b\c{CPU P2} Same as 686
4689 \b\c{CPU P3} Pentium III (Katmai) instruction sets
4691 \b\c{CPU KATMAI} Same as P3
4693 \b\c{CPU P4} Pentium 4 (Willamette) instruction set
4695 \b\c{CPU WILLAMETTE} Same as P4
4697 \b\c{CPU PRESCOTT} Prescott instruction set
4699 \b\c{CPU X64} x86-64 (x64/AMD64/Intel 64) instruction set
4701 \b\c{CPU IA64} IA64 CPU (in x86 mode) instruction set
4703 All options are case insensitive. All instructions will be selected
4704 only if they apply to the selected CPU or lower. By default, all
4705 instructions are available.
4708 \H{FLOAT} \i\c{FLOAT}: Handling of \I{floating-point, constants}floating-point constants
4710 By default, floating-point constants are rounded to nearest, and IEEE
4711 denormals are supported. The following options can be set to alter
4714 \b\c{FLOAT DAZ} Flush denormals to zero
4716 \b\c{FLOAT NODAZ} Do not flush denormals to zero (default)
4718 \b\c{FLOAT NEAR} Round to nearest (default)
4720 \b\c{FLOAT UP} Round up (toward +Infinity)
4722 \b\c{FLOAT DOWN} Round down (toward -Infinity)
4724 \b\c{FLOAT ZERO} Round toward zero
4726 \b\c{FLOAT DEFAULT} Restore default settings
4728 The standard macros \i\c{__FLOAT_DAZ__}, \i\c{__FLOAT_ROUND__}, and
4729 \i\c{__FLOAT__} contain the current state, as long as the programmer
4730 has avoided the use of the brackeded primitive form, (\c{[FLOAT]}).
4732 \c{__FLOAT__} contains the full set of floating-point settings; this
4733 value can be saved away and invoked later to restore the setting.
4736 \C{outfmt} \i{Output Formats}
4738 NASM is a portable assembler, designed to be able to compile on any
4739 ANSI C-supporting platform and produce output to run on a variety of
4740 Intel x86 operating systems. For this reason, it has a large number
4741 of available output formats, selected using the \i\c{-f} option on
4742 the NASM \i{command line}. Each of these formats, along with its
4743 extensions to the base NASM syntax, is detailed in this chapter.
4745 As stated in \k{opt-o}, NASM chooses a \i{default name} for your
4746 output file based on the input file name and the chosen output
4747 format. This will be generated by removing the \i{extension}
4748 (\c{.asm}, \c{.s}, or whatever you like to use) from the input file
4749 name, and substituting an extension defined by the output format.
4750 The extensions are given with each format below.
4753 \H{binfmt} \i\c{bin}: \i{Flat-Form Binary}\I{pure binary} Output
4755 The \c{bin} format does not produce object files: it generates
4756 nothing in the output file except the code you wrote. Such `pure
4757 binary' files are used by \i{MS-DOS}: \i\c{.COM} executables and
4758 \i\c{.SYS} device drivers are pure binary files. Pure binary output
4759 is also useful for \i{operating system} and \i{boot loader}
4762 The \c{bin} format supports \i{multiple section names}. For details of
4763 how NASM handles sections in the \c{bin} format, see \k{multisec}.
4765 Using the \c{bin} format puts NASM by default into 16-bit mode (see
4766 \k{bits}). In order to use \c{bin} to write 32-bit or 64-bit code,
4767 such as an OS kernel, you need to explicitly issue the \I\c{BITS}\c{BITS 32}
4768 or \I\c{BITS}\c{BITS 64} directive.
4770 \c{bin} has no default output file name extension: instead, it
4771 leaves your file name as it is once the original extension has been
4772 removed. Thus, the default is for NASM to assemble \c{binprog.asm}
4773 into a binary file called \c{binprog}.
4776 \S{org} \i\c{ORG}: Binary File \i{Program Origin}
4778 The \c{bin} format provides an additional directive to the list
4779 given in \k{directive}: \c{ORG}. The function of the \c{ORG}
4780 directive is to specify the origin address which NASM will assume
4781 the program begins at when it is loaded into memory.
4783 For example, the following code will generate the longword
4790 Unlike the \c{ORG} directive provided by MASM-compatible assemblers,
4791 which allows you to jump around in the object file and overwrite
4792 code you have already generated, NASM's \c{ORG} does exactly what
4793 the directive says: \e{origin}. Its sole function is to specify one
4794 offset which is added to all internal address references within the
4795 section; it does not permit any of the trickery that MASM's version
4796 does. See \k{proborg} for further comments.
4799 \S{binseg} \c{bin} Extensions to the \c{SECTION}
4800 Directive\I{SECTION, bin extensions to}
4802 The \c{bin} output format extends the \c{SECTION} (or \c{SEGMENT})
4803 directive to allow you to specify the alignment requirements of
4804 segments. This is done by appending the \i\c{ALIGN} qualifier to the
4805 end of the section-definition line. For example,
4807 \c section .data align=16
4809 switches to the section \c{.data} and also specifies that it must be
4810 aligned on a 16-byte boundary.
4812 The parameter to \c{ALIGN} specifies how many low bits of the
4813 section start address must be forced to zero. The alignment value
4814 given may be any power of two.\I{section alignment, in
4815 bin}\I{segment alignment, in bin}\I{alignment, in bin sections}
4818 \S{multisec} \i{Multisection}\I{bin, multisection} Support for the \c{bin} Format
4820 The \c{bin} format allows the use of multiple sections, of arbitrary names,
4821 besides the "known" \c{.text}, \c{.data}, and \c{.bss} names.
4823 \b Sections may be designated \i\c{progbits} or \i\c{nobits}. Default
4824 is \c{progbits} (except \c{.bss}, which defaults to \c{nobits},
4827 \b Sections can be aligned at a specified boundary following the previous
4828 section with \c{align=}, or at an arbitrary byte-granular position with
4831 \b Sections can be given a virtual start address, which will be used
4832 for the calculation of all memory references within that section
4835 \b Sections can be ordered using \i\c{follows=}\c{<section>} or
4836 \i\c{vfollows=}\c{<section>} as an alternative to specifying an explicit
4839 \b Arguments to \c{org}, \c{start}, \c{vstart}, and \c{align=} are
4840 critical expressions. See \k{crit}. E.g. \c{align=(1 << ALIGN_SHIFT)}
4841 - \c{ALIGN_SHIFT} must be defined before it is used here.
4843 \b Any code which comes before an explicit \c{SECTION} directive
4844 is directed by default into the \c{.text} section.
4846 \b If an \c{ORG} statement is not given, \c{ORG 0} is used
4849 \b The \c{.bss} section will be placed after the last \c{progbits}
4850 section, unless \c{start=}, \c{vstart=}, \c{follows=}, or \c{vfollows=}
4853 \b All sections are aligned on dword boundaries, unless a different
4854 alignment has been specified.
4856 \b Sections may not overlap.
4858 \b NASM creates the \c{section.<secname>.start} for each section,
4859 which may be used in your code.
4861 \S{map}\i{Map Files}
4863 Map files can be generated in \c{-f bin} format by means of the \c{[map]}
4864 option. Map types of \c{all} (default), \c{brief}, \c{sections}, \c{segments},
4865 or \c{symbols} may be specified. Output may be directed to \c{stdout}
4866 (default), \c{stderr}, or a specified file. E.g.
4867 \c{[map symbols myfile.map]}. No "user form" exists, the square
4868 brackets must be used.
4871 \H{ithfmt} \i\c{ith}: \i{Intel Hex} Output
4873 The \c{ith} file format produces Intel hex-format files. Just as the
4874 \c{bin} format, this is a flat memory image format with no support for
4875 relocation or linking. It is usually used with ROM programmers and
4878 All extensions supported by the \c{bin} file format is also supported by
4879 the \c{ith} file format.
4881 \c{ith} provides a default output file-name extension of \c{.ith}.
4884 \H{srecfmt} \i\c{srec}: \i{Motorola S-Records} Output
4886 The \c{srec} file format produces Motorola S-records files. Just as the
4887 \c{bin} format, this is a flat memory image format with no support for
4888 relocation or linking. It is usually used with ROM programmers and
4891 All extensions supported by the \c{bin} file format is also supported by
4892 the \c{srec} file format.
4894 \c{srec} provides a default output file-name extension of \c{.srec}.
4897 \H{objfmt} \i\c{obj}: \i{Microsoft OMF}\I{OMF} Object Files
4899 The \c{obj} file format (NASM calls it \c{obj} rather than \c{omf}
4900 for historical reasons) is the one produced by \i{MASM} and
4901 \i{TASM}, which is typically fed to 16-bit DOS linkers to produce
4902 \i\c{.EXE} files. It is also the format used by \i{OS/2}.
4904 \c{obj} provides a default output file-name extension of \c{.obj}.
4906 \c{obj} is not exclusively a 16-bit format, though: NASM has full
4907 support for the 32-bit extensions to the format. In particular,
4908 32-bit \c{obj} format files are used by \i{Borland's Win32
4909 compilers}, instead of using Microsoft's newer \i\c{win32} object
4912 The \c{obj} format does not define any special segment names: you
4913 can call your segments anything you like. Typical names for segments
4914 in \c{obj} format files are \c{CODE}, \c{DATA} and \c{BSS}.
4916 If your source file contains code before specifying an explicit
4917 \c{SEGMENT} directive, then NASM will invent its own segment called
4918 \i\c{__NASMDEFSEG} for you.
4920 When you define a segment in an \c{obj} file, NASM defines the
4921 segment name as a symbol as well, so that you can access the segment
4922 address of the segment. So, for example:
4931 \c mov ax,data ; get segment address of data
4932 \c mov ds,ax ; and move it into DS
4933 \c inc word [dvar] ; now this reference will work
4936 The \c{obj} format also enables the use of the \i\c{SEG} and
4937 \i\c{WRT} operators, so that you can write code which does things
4942 \c mov ax,seg foo ; get preferred segment of foo
4944 \c mov ax,data ; a different segment
4946 \c mov ax,[ds:foo] ; this accesses `foo'
4947 \c mov [es:foo wrt data],bx ; so does this
4950 \S{objseg} \c{obj} Extensions to the \c{SEGMENT}
4951 Directive\I{SEGMENT, obj extensions to}
4953 The \c{obj} output format extends the \c{SEGMENT} (or \c{SECTION})
4954 directive to allow you to specify various properties of the segment
4955 you are defining. This is done by appending extra qualifiers to the
4956 end of the segment-definition line. For example,
4958 \c segment code private align=16
4960 defines the segment \c{code}, but also declares it to be a private
4961 segment, and requires that the portion of it described in this code
4962 module must be aligned on a 16-byte boundary.
4964 The available qualifiers are:
4966 \b \i\c{PRIVATE}, \i\c{PUBLIC}, \i\c{COMMON} and \i\c{STACK} specify
4967 the combination characteristics of the segment. \c{PRIVATE} segments
4968 do not get combined with any others by the linker; \c{PUBLIC} and
4969 \c{STACK} segments get concatenated together at link time; and
4970 \c{COMMON} segments all get overlaid on top of each other rather
4971 than stuck end-to-end.
4973 \b \i\c{ALIGN} is used, as shown above, to specify how many low bits
4974 of the segment start address must be forced to zero. The alignment
4975 value given may be any power of two from 1 to 4096; in reality, the
4976 only values supported are 1, 2, 4, 16, 256 and 4096, so if 8 is
4977 specified it will be rounded up to 16, and 32, 64 and 128 will all
4978 be rounded up to 256, and so on. Note that alignment to 4096-byte
4979 boundaries is a \i{PharLap} extension to the format and may not be
4980 supported by all linkers.\I{section alignment, in OBJ}\I{segment
4981 alignment, in OBJ}\I{alignment, in OBJ sections}
4983 \b \i\c{CLASS} can be used to specify the segment class; this feature
4984 indicates to the linker that segments of the same class should be
4985 placed near each other in the output file. The class name can be any
4986 word, e.g. \c{CLASS=CODE}.
4988 \b \i\c{OVERLAY}, like \c{CLASS}, is specified with an arbitrary word
4989 as an argument, and provides overlay information to an
4990 overlay-capable linker.
4992 \b Segments can be declared as \i\c{USE16} or \i\c{USE32}, which has
4993 the effect of recording the choice in the object file and also
4994 ensuring that NASM's default assembly mode when assembling in that
4995 segment is 16-bit or 32-bit respectively.
4997 \b When writing \i{OS/2} object files, you should declare 32-bit
4998 segments as \i\c{FLAT}, which causes the default segment base for
4999 anything in the segment to be the special group \c{FLAT}, and also
5000 defines the group if it is not already defined.
5002 \b The \c{obj} file format also allows segments to be declared as
5003 having a pre-defined absolute segment address, although no linkers
5004 are currently known to make sensible use of this feature;
5005 nevertheless, NASM allows you to declare a segment such as
5006 \c{SEGMENT SCREEN ABSOLUTE=0xB800} if you need to. The \i\c{ABSOLUTE}
5007 and \c{ALIGN} keywords are mutually exclusive.
5009 NASM's default segment attributes are \c{PUBLIC}, \c{ALIGN=1}, no
5010 class, no overlay, and \c{USE16}.
5013 \S{group} \i\c{GROUP}: Defining Groups of Segments\I{segments, groups of}
5015 The \c{obj} format also allows segments to be grouped, so that a
5016 single segment register can be used to refer to all the segments in
5017 a group. NASM therefore supplies the \c{GROUP} directive, whereby
5026 \c ; some uninitialized data
5028 \c group dgroup data bss
5030 which will define a group called \c{dgroup} to contain the segments
5031 \c{data} and \c{bss}. Like \c{SEGMENT}, \c{GROUP} causes the group
5032 name to be defined as a symbol, so that you can refer to a variable
5033 \c{var} in the \c{data} segment as \c{var wrt data} or as \c{var wrt
5034 dgroup}, depending on which segment value is currently in your
5037 If you just refer to \c{var}, however, and \c{var} is declared in a
5038 segment which is part of a group, then NASM will default to giving
5039 you the offset of \c{var} from the beginning of the \e{group}, not
5040 the \e{segment}. Therefore \c{SEG var}, also, will return the group
5041 base rather than the segment base.
5043 NASM will allow a segment to be part of more than one group, but
5044 will generate a warning if you do this. Variables declared in a
5045 segment which is part of more than one group will default to being
5046 relative to the first group that was defined to contain the segment.
5048 A group does not have to contain any segments; you can still make
5049 \c{WRT} references to a group which does not contain the variable
5050 you are referring to. OS/2, for example, defines the special group
5051 \c{FLAT} with no segments in it.
5054 \S{uppercase} \i\c{UPPERCASE}: Disabling Case Sensitivity in Output
5056 Although NASM itself is \i{case sensitive}, some OMF linkers are
5057 not; therefore it can be useful for NASM to output single-case
5058 object files. The \c{UPPERCASE} format-specific directive causes all
5059 segment, group and symbol names that are written to the object file
5060 to be forced to upper case just before being written. Within a
5061 source file, NASM is still case-sensitive; but the object file can
5062 be written entirely in upper case if desired.
5064 \c{UPPERCASE} is used alone on a line; it requires no parameters.
5067 \S{import} \i\c{IMPORT}: Importing DLL Symbols\I{DLL symbols,
5068 importing}\I{symbols, importing from DLLs}
5070 The \c{IMPORT} format-specific directive defines a symbol to be
5071 imported from a DLL, for use if you are writing a DLL's \i{import
5072 library} in NASM. You still need to declare the symbol as \c{EXTERN}
5073 as well as using the \c{IMPORT} directive.
5075 The \c{IMPORT} directive takes two required parameters, separated by
5076 white space, which are (respectively) the name of the symbol you
5077 wish to import and the name of the library you wish to import it
5080 \c import WSAStartup wsock32.dll
5082 A third optional parameter gives the name by which the symbol is
5083 known in the library you are importing it from, in case this is not
5084 the same as the name you wish the symbol to be known by to your code
5085 once you have imported it. For example:
5087 \c import asyncsel wsock32.dll WSAAsyncSelect
5090 \S{export} \i\c{EXPORT}: Exporting DLL Symbols\I{DLL symbols,
5091 exporting}\I{symbols, exporting from DLLs}
5093 The \c{EXPORT} format-specific directive defines a global symbol to
5094 be exported as a DLL symbol, for use if you are writing a DLL in
5095 NASM. You still need to declare the symbol as \c{GLOBAL} as well as
5096 using the \c{EXPORT} directive.
5098 \c{EXPORT} takes one required parameter, which is the name of the
5099 symbol you wish to export, as it was defined in your source file. An
5100 optional second parameter (separated by white space from the first)
5101 gives the \e{external} name of the symbol: the name by which you
5102 wish the symbol to be known to programs using the DLL. If this name
5103 is the same as the internal name, you may leave the second parameter
5106 Further parameters can be given to define attributes of the exported
5107 symbol. These parameters, like the second, are separated by white
5108 space. If further parameters are given, the external name must also
5109 be specified, even if it is the same as the internal name. The
5110 available attributes are:
5112 \b \c{resident} indicates that the exported name is to be kept
5113 resident by the system loader. This is an optimisation for
5114 frequently used symbols imported by name.
5116 \b \c{nodata} indicates that the exported symbol is a function which
5117 does not make use of any initialized data.
5119 \b \c{parm=NNN}, where \c{NNN} is an integer, sets the number of
5120 parameter words for the case in which the symbol is a call gate
5121 between 32-bit and 16-bit segments.
5123 \b An attribute which is just a number indicates that the symbol
5124 should be exported with an identifying number (ordinal), and gives
5130 \c export myfunc TheRealMoreFormalLookingFunctionName
5131 \c export myfunc myfunc 1234 ; export by ordinal
5132 \c export myfunc myfunc resident parm=23 nodata
5135 \S{dotdotstart} \i\c{..start}: Defining the \i{Program Entry
5138 \c{OMF} linkers require exactly one of the object files being linked to
5139 define the program entry point, where execution will begin when the
5140 program is run. If the object file that defines the entry point is
5141 assembled using NASM, you specify the entry point by declaring the
5142 special symbol \c{..start} at the point where you wish execution to
5146 \S{objextern} \c{obj} Extensions to the \c{EXTERN}
5147 Directive\I{EXTERN, obj extensions to}
5149 If you declare an external symbol with the directive
5153 then references such as \c{mov ax,foo} will give you the offset of
5154 \c{foo} from its preferred segment base (as specified in whichever
5155 module \c{foo} is actually defined in). So to access the contents of
5156 \c{foo} you will usually need to do something like
5158 \c mov ax,seg foo ; get preferred segment base
5159 \c mov es,ax ; move it into ES
5160 \c mov ax,[es:foo] ; and use offset `foo' from it
5162 This is a little unwieldy, particularly if you know that an external
5163 is going to be accessible from a given segment or group, say
5164 \c{dgroup}. So if \c{DS} already contained \c{dgroup}, you could
5167 \c mov ax,[foo wrt dgroup]
5169 However, having to type this every time you want to access \c{foo}
5170 can be a pain; so NASM allows you to declare \c{foo} in the
5173 \c extern foo:wrt dgroup
5175 This form causes NASM to pretend that the preferred segment base of
5176 \c{foo} is in fact \c{dgroup}; so the expression \c{seg foo} will
5177 now return \c{dgroup}, and the expression \c{foo} is equivalent to
5180 This \I{default-WRT mechanism}default-\c{WRT} mechanism can be used
5181 to make externals appear to be relative to any group or segment in
5182 your program. It can also be applied to common variables: see
5186 \S{objcommon} \c{obj} Extensions to the \c{COMMON}
5187 Directive\I{COMMON, obj extensions to}
5189 The \c{obj} format allows common variables to be either near\I{near
5190 common variables} or far\I{far common variables}; NASM allows you to
5191 specify which your variables should be by the use of the syntax
5193 \c common nearvar 2:near ; `nearvar' is a near common
5194 \c common farvar 10:far ; and `farvar' is far
5196 Far common variables may be greater in size than 64Kb, and so the
5197 OMF specification says that they are declared as a number of
5198 \e{elements} of a given size. So a 10-byte far common variable could
5199 be declared as ten one-byte elements, five two-byte elements, two
5200 five-byte elements or one ten-byte element.
5202 Some \c{OMF} linkers require the \I{element size, in common
5203 variables}\I{common variables, element size}element size, as well as
5204 the variable size, to match when resolving common variables declared
5205 in more than one module. Therefore NASM must allow you to specify
5206 the element size on your far common variables. This is done by the
5209 \c common c_5by2 10:far 5 ; two five-byte elements
5210 \c common c_2by5 10:far 2 ; five two-byte elements
5212 If no element size is specified, the default is 1. Also, the \c{FAR}
5213 keyword is not required when an element size is specified, since
5214 only far commons may have element sizes at all. So the above
5215 declarations could equivalently be
5217 \c common c_5by2 10:5 ; two five-byte elements
5218 \c common c_2by5 10:2 ; five two-byte elements
5220 In addition to these extensions, the \c{COMMON} directive in \c{obj}
5221 also supports default-\c{WRT} specification like \c{EXTERN} does
5222 (explained in \k{objextern}). So you can also declare things like
5224 \c common foo 10:wrt dgroup
5225 \c common bar 16:far 2:wrt data
5226 \c common baz 24:wrt data:6
5229 \H{win32fmt} \i\c{win32}: Microsoft Win32 Object Files
5231 The \c{win32} output format generates Microsoft Win32 object files,
5232 suitable for passing to Microsoft linkers such as \i{Visual C++}.
5233 Note that Borland Win32 compilers do not use this format, but use
5234 \c{obj} instead (see \k{objfmt}).
5236 \c{win32} provides a default output file-name extension of \c{.obj}.
5238 Note that although Microsoft say that Win32 object files follow the
5239 \c{COFF} (Common Object File Format) standard, the object files produced
5240 by Microsoft Win32 compilers are not compatible with COFF linkers
5241 such as DJGPP's, and vice versa. This is due to a difference of
5242 opinion over the precise semantics of PC-relative relocations. To
5243 produce COFF files suitable for DJGPP, use NASM's \c{coff} output
5244 format; conversely, the \c{coff} format does not produce object
5245 files that Win32 linkers can generate correct output from.
5248 \S{win32sect} \c{win32} Extensions to the \c{SECTION}
5249 Directive\I{SECTION, win32 extensions to}
5251 Like the \c{obj} format, \c{win32} allows you to specify additional
5252 information on the \c{SECTION} directive line, to control the type
5253 and properties of sections you declare. Section types and properties
5254 are generated automatically by NASM for the \i{standard section names}
5255 \c{.text}, \c{.data} and \c{.bss}, but may still be overridden by
5258 The available qualifiers are:
5260 \b \c{code}, or equivalently \c{text}, defines the section to be a
5261 code section. This marks the section as readable and executable, but
5262 not writable, and also indicates to the linker that the type of the
5265 \b \c{data} and \c{bss} define the section to be a data section,
5266 analogously to \c{code}. Data sections are marked as readable and
5267 writable, but not executable. \c{data} declares an initialized data
5268 section, whereas \c{bss} declares an uninitialized data section.
5270 \b \c{rdata} declares an initialized data section that is readable
5271 but not writable. Microsoft compilers use this section to place
5274 \b \c{info} defines the section to be an \i{informational section},
5275 which is not included in the executable file by the linker, but may
5276 (for example) pass information \e{to} the linker. For example,
5277 declaring an \c{info}-type section called \i\c{.drectve} causes the
5278 linker to interpret the contents of the section as command-line
5281 \b \c{align=}, used with a trailing number as in \c{obj}, gives the
5282 \I{section alignment, in win32}\I{alignment, in win32
5283 sections}alignment requirements of the section. The maximum you may
5284 specify is 64: the Win32 object file format contains no means to
5285 request a greater section alignment than this. If alignment is not
5286 explicitly specified, the defaults are 16-byte alignment for code
5287 sections, 8-byte alignment for rdata sections and 4-byte alignment
5288 for data (and BSS) sections.
5289 Informational sections get a default alignment of 1 byte (no
5290 alignment), though the value does not matter.
5292 The defaults assumed by NASM if you do not specify the above
5295 \c section .text code align=16
5296 \c section .data data align=4
5297 \c section .rdata rdata align=8
5298 \c section .bss bss align=4
5300 Any other section name is treated by default like \c{.text}.
5302 \S{win32safeseh} \c{win32}: Safe Structured Exception Handling
5304 Among other improvements in Windows XP SP2 and Windows Server 2003
5305 Microsoft has introduced concept of "safe structured exception
5306 handling." General idea is to collect handlers' entry points in
5307 designated read-only table and have alleged entry point verified
5308 against this table prior exception control is passed to the handler. In
5309 order for an executable module to be equipped with such "safe exception
5310 handler table," all object modules on linker command line has to comply
5311 with certain criteria. If one single module among them does not, then
5312 the table in question is omitted and above mentioned run-time checks
5313 will not be performed for application in question. Table omission is by
5314 default silent and therefore can be easily overlooked. One can instruct
5315 linker to refuse to produce binary without such table by passing
5316 \c{/safeseh} command line option.
5318 Without regard to this run-time check merits it's natural to expect
5319 NASM to be capable of generating modules suitable for \c{/safeseh}
5320 linking. From developer's viewpoint the problem is two-fold:
5322 \b how to adapt modules not deploying exception handlers of their own;
5324 \b how to adapt/develop modules utilizing custom exception handling;
5326 Former can be easily achieved with any NASM version by adding following
5327 line to source code:
5331 As of version 2.03 NASM adds this absolute symbol automatically. If
5332 it's not already present to be precise. I.e. if for whatever reason
5333 developer would choose to assign another value in source file, it would
5334 still be perfectly possible.
5336 Registering custom exception handler on the other hand requires certain
5337 "magic." As of version 2.03 additional directive is implemented,
5338 \c{safeseh}, which instructs the assembler to produce appropriately
5339 formatted input data for above mentioned "safe exception handler
5340 table." Its typical use would be:
5343 \c extern _MessageBoxA@16
5344 \c %if __NASM_VERSION_ID__ >= 0x02030000
5345 \c safeseh handler ; register handler as "safe handler"
5348 \c push DWORD 1 ; MB_OKCANCEL
5349 \c push DWORD caption
5352 \c call _MessageBoxA@16
5353 \c sub eax,1 ; incidentally suits as return value
5354 \c ; for exception handler
5358 \c push DWORD handler
5359 \c push DWORD [fs:0]
5360 \c mov DWORD [fs:0],esp ; engage exception handler
5362 \c mov eax,DWORD[eax] ; cause exception
5363 \c pop DWORD [fs:0] ; disengage exception handler
5366 \c text: db 'OK to rethrow, CANCEL to generate core dump',0
5367 \c caption:db 'SEGV',0
5369 \c section .drectve info
5370 \c db '/defaultlib:user32.lib /defaultlib:msvcrt.lib '
5372 As you might imagine, it's perfectly possible to produce .exe binary
5373 with "safe exception handler table" and yet engage unregistered
5374 exception handler. Indeed, handler is engaged by simply manipulating
5375 \c{[fs:0]} location at run-time, something linker has no power over,
5376 run-time that is. It should be explicitly mentioned that such failure
5377 to register handler's entry point with \c{safeseh} directive has
5378 undesired side effect at run-time. If exception is raised and
5379 unregistered handler is to be executed, the application is abruptly
5380 terminated without any notification whatsoever. One can argue that
5381 system could at least have logged some kind "non-safe exception
5382 handler in x.exe at address n" message in event log, but no, literally
5383 no notification is provided and user is left with no clue on what
5384 caused application failure.
5386 Finally, all mentions of linker in this paragraph refer to Microsoft
5387 linker version 7.x and later. Presence of \c{@feat.00} symbol and input
5388 data for "safe exception handler table" causes no backward
5389 incompatibilities and "safeseh" modules generated by NASM 2.03 and
5390 later can still be linked by earlier versions or non-Microsoft linkers.
5393 \H{win64fmt} \i\c{win64}: Microsoft Win64 Object Files
5395 The \c{win64} output format generates Microsoft Win64 object files,
5396 which is nearly 100% identical to the \c{win32} object format (\k{win32fmt})
5397 with the exception that it is meant to target 64-bit code and the x86-64
5398 platform altogether. This object file is used exactly the same as the \c{win32}
5399 object format (\k{win32fmt}), in NASM, with regard to this exception.
5401 \S{win64pic} \c{win64}: Writing Position-Independent Code
5403 While \c{REL} takes good care of RIP-relative addressing, there is one
5404 aspect that is easy to overlook for a Win64 programmer: indirect
5405 references. Consider a switch dispatch table:
5407 \c jmp QWORD[dsptch+rax*8]
5413 Even novice Win64 assembler programmer will soon realize that the code
5414 is not 64-bit savvy. Most notably linker will refuse to link it with
5415 "\c{'ADDR32' relocation to '.text' invalid without
5416 /LARGEADDRESSAWARE:NO}". So [s]he will have to split jmp instruction as
5419 \c lea rbx,[rel dsptch]
5420 \c jmp QWORD[rbx+rax*8]
5422 What happens behind the scene is that effective address in \c{lea} is
5423 encoded relative to instruction pointer, or in perfectly
5424 position-independent manner. But this is only part of the problem!
5425 Trouble is that in .dll context \c{caseN} relocations will make their
5426 way to the final module and might have to be adjusted at .dll load
5427 time. To be specific when it can't be loaded at preferred address. And
5428 when this occurs, pages with such relocations will be rendered private
5429 to current process, which kind of undermines the idea of sharing .dll.
5430 But no worry, it's trivial to fix:
5432 \c lea rbx,[rel dsptch]
5433 \c add rbx,QWORD[rbx+rax*8]
5436 \c dsptch: dq case0-dsptch
5440 NASM version 2.03 and later provides another alternative, \c{wrt
5441 ..imagebase} operator, which returns offset from base address of the
5442 current image, be it .exe or .dll module, therefore the name. For those
5443 acquainted with PE-COFF format base address denotes start of
5444 \c{IMAGE_DOS_HEADER} structure. Here is how to implement switch with
5445 these image-relative references:
5447 \c lea rbx,[rel dsptch]
5448 \c mov eax,DWORD[rbx+rax*4]
5449 \c sub rbx,dsptch wrt ..imagebase
5453 \c dsptch: dd case0 wrt ..imagebase
5454 \c dd case1 wrt ..imagebase
5456 One can argue that the operator is redundant. Indeed, snippet before
5457 last works just fine with any NASM version and is not even Windows
5458 specific... The real reason for implementing \c{wrt ..imagebase} will
5459 become apparent in next paragraph.
5461 It should be noted that \c{wrt ..imagebase} is defined as 32-bit
5464 \c dd label wrt ..imagebase ; ok
5465 \c dq label wrt ..imagebase ; bad
5466 \c mov eax,label wrt ..imagebase ; ok
5467 \c mov rax,label wrt ..imagebase ; bad
5469 \S{win64seh} \c{win64}: Structured Exception Handling
5471 Structured exception handing in Win64 is completely different matter
5472 from Win32. Upon exception program counter value is noted, and
5473 linker-generated table comprising start and end addresses of all the
5474 functions [in given executable module] is traversed and compared to the
5475 saved program counter. Thus so called \c{UNWIND_INFO} structure is
5476 identified. If it's not found, then offending subroutine is assumed to
5477 be "leaf" and just mentioned lookup procedure is attempted for its
5478 caller. In Win64 leaf function is such function that does not call any
5479 other function \e{nor} modifies any Win64 non-volatile registers,
5480 including stack pointer. The latter ensures that it's possible to
5481 identify leaf function's caller by simply pulling the value from the
5484 While majority of subroutines written in assembler are not calling any
5485 other function, requirement for non-volatile registers' immutability
5486 leaves developer with not more than 7 registers and no stack frame,
5487 which is not necessarily what [s]he counted with. Customarily one would
5488 meet the requirement by saving non-volatile registers on stack and
5489 restoring them upon return, so what can go wrong? If [and only if] an
5490 exception is raised at run-time and no \c{UNWIND_INFO} structure is
5491 associated with such "leaf" function, the stack unwind procedure will
5492 expect to find caller's return address on the top of stack immediately
5493 followed by its frame. Given that developer pushed caller's
5494 non-volatile registers on stack, would the value on top point at some
5495 code segment or even addressable space? Well, developer can attempt
5496 copying caller's return address to the top of stack and this would
5497 actually work in some very specific circumstances. But unless developer
5498 can guarantee that these circumstances are always met, it's more
5499 appropriate to assume worst case scenario, i.e. stack unwind procedure
5500 going berserk. Relevant question is what happens then? Application is
5501 abruptly terminated without any notification whatsoever. Just like in
5502 Win32 case, one can argue that system could at least have logged
5503 "unwind procedure went berserk in x.exe at address n" in event log, but
5504 no, no trace of failure is left.
5506 Now, when we understand significance of the \c{UNWIND_INFO} structure,
5507 let's discuss what's in it and/or how it's processed. First of all it
5508 is checked for presence of reference to custom language-specific
5509 exception handler. If there is one, then it's invoked. Depending on the
5510 return value, execution flow is resumed (exception is said to be
5511 "handled"), \e{or} rest of \c{UNWIND_INFO} structure is processed as
5512 following. Beside optional reference to custom handler, it carries
5513 information about current callee's stack frame and where non-volatile
5514 registers are saved. Information is detailed enough to be able to
5515 reconstruct contents of caller's non-volatile registers upon call to
5516 current callee. And so caller's context is reconstructed, and then
5517 unwind procedure is repeated, i.e. another \c{UNWIND_INFO} structure is
5518 associated, this time, with caller's instruction pointer, which is then
5519 checked for presence of reference to language-specific handler, etc.
5520 The procedure is recursively repeated till exception is handled. As
5521 last resort system "handles" it by generating memory core dump and
5522 terminating the application.
5524 As for the moment of this writing NASM unfortunately does not
5525 facilitate generation of above mentioned detailed information about
5526 stack frame layout. But as of version 2.03 it implements building
5527 blocks for generating structures involved in stack unwinding. As
5528 simplest example, here is how to deploy custom exception handler for
5533 \c extern MessageBoxA
5539 \c mov r9,1 ; MB_OKCANCEL
5541 \c sub eax,1 ; incidentally suits as return value
5542 \c ; for exception handler
5548 \c mov rax,QWORD[rax] ; cause exception
5551 \c text: db 'OK to rethrow, CANCEL to generate core dump',0
5552 \c caption:db 'SEGV',0
5554 \c section .pdata rdata align=4
5555 \c dd main wrt ..imagebase
5556 \c dd main_end wrt ..imagebase
5557 \c dd xmain wrt ..imagebase
5558 \c section .xdata rdata align=8
5559 \c xmain: db 9,0,0,0
5560 \c dd handler wrt ..imagebase
5561 \c section .drectve info
5562 \c db '/defaultlib:user32.lib /defaultlib:msvcrt.lib '
5564 What you see in \c{.pdata} section is element of the "table comprising
5565 start and end addresses of function" along with reference to associated
5566 \c{UNWIND_INFO} structure. And what you see in \c{.xdata} section is
5567 \c{UNWIND_INFO} structure describing function with no frame, but with
5568 designated exception handler. References are \e{required} to be
5569 image-relative (which is the real reason for implementing \c{wrt
5570 ..imagebase} operator). It should be noted that \c{rdata align=n}, as
5571 well as \c{wrt ..imagebase}, are optional in these two segments'
5572 contexts, i.e. can be omitted. Latter means that \e{all} 32-bit
5573 references, not only above listed required ones, placed into these two
5574 segments turn out image-relative. Why is it important to understand?
5575 Developer is allowed to append handler-specific data to \c{UNWIND_INFO}
5576 structure, and if [s]he adds a 32-bit reference, then [s]he will have
5577 to remember to adjust its value to obtain the real pointer.
5579 As already mentioned, in Win64 terms leaf function is one that does not
5580 call any other function \e{nor} modifies any non-volatile register,
5581 including stack pointer. But it's not uncommon that assembler
5582 programmer plans to utilize every single register and sometimes even
5583 have variable stack frame. Is there anything one can do with bare
5584 building blocks? I.e. besides manually composing fully-fledged
5585 \c{UNWIND_INFO} structure, which would surely be considered
5586 error-prone? Yes, there is. Recall that exception handler is called
5587 first, before stack layout is analyzed. As it turned out, it's
5588 perfectly possible to manipulate current callee's context in custom
5589 handler in manner that permits further stack unwinding. General idea is
5590 that handler would not actually "handle" the exception, but instead
5591 restore callee's context, as it was at its entry point and thus mimic
5592 leaf function. In other words, handler would simply undertake part of
5593 unwinding procedure. Consider following example:
5596 \c mov rax,rsp ; copy rsp to volatile register
5597 \c push r15 ; save non-volatile registers
5600 \c mov r11,rsp ; prepare variable stack frame
5603 \c mov QWORD[r11],rax ; check for exceptions
5604 \c mov rsp,r11 ; allocate stack frame
5605 \c mov QWORD[rsp],rax ; save original rsp value
5608 \c mov r11,QWORD[rsp] ; pull original rsp value
5609 \c mov rbp,QWORD[r11-24]
5610 \c mov rbx,QWORD[r11-16]
5611 \c mov r15,QWORD[r11-8]
5612 \c mov rsp,r11 ; destroy frame
5615 The keyword is that up to \c{magic_point} original \c{rsp} value
5616 remains in chosen volatile register and no non-volatile register,
5617 except for \c{rsp}, is modified. While past \c{magic_point} \c{rsp}
5618 remains constant till the very end of the \c{function}. In this case
5619 custom language-specific exception handler would look like this:
5621 \c EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
5622 \c CONTEXT *context,DISPATCHER_CONTEXT *disp)
5624 \c if (context->Rip<(ULONG64)magic_point)
5625 \c rsp = (ULONG64 *)context->Rax;
5627 \c { rsp = ((ULONG64 **)context->Rsp)[0];
5628 \c context->Rbp = rsp[-3];
5629 \c context->Rbx = rsp[-2];
5630 \c context->R15 = rsp[-1];
5632 \c context->Rsp = (ULONG64)rsp;
5634 \c memcpy (disp->ContextRecord,context,sizeof(CONTEXT));
5635 \c RtlVirtualUnwind(UNW_FLAG_NHANDLER,disp->ImageBase,
5636 \c dips->ControlPc,disp->FunctionEntry,disp->ContextRecord,
5637 \c &disp->HandlerData,&disp->EstablisherFrame,NULL);
5638 \c return ExceptionContinueSearch;
5641 As custom handler mimics leaf function, corresponding \c{UNWIND_INFO}
5642 structure does not have to contain any information about stack frame
5645 \H{cofffmt} \i\c{coff}: \i{Common Object File Format}
5647 The \c{coff} output type produces \c{COFF} object files suitable for
5648 linking with the \i{DJGPP} linker.
5650 \c{coff} provides a default output file-name extension of \c{.o}.
5652 The \c{coff} format supports the same extensions to the \c{SECTION}
5653 directive as \c{win32} does, except that the \c{align} qualifier and
5654 the \c{info} section type are not supported.
5656 \H{machofmt} \I{Mach-O}\i\c{macho32} and \i\c{macho64}: \i{Mach Object File Format}
5658 The \c{macho32} and \c{macho64} output formts produces \c{Mach-O}
5659 object files suitable for linking with the \i{MacOS X} linker.
5660 \i\c{macho} is a synonym for \c{macho32}.
5662 \c{macho} provides a default output file-name extension of \c{.o}.
5664 \H{elffmt} \i\c{elf32} and \i\c{elf64}: \I{ELF}\I{linux, elf}\i{Executable and Linkable
5665 Format} Object Files
5667 The \c{elf32} and \c{elf64} output formats generate \c{ELF32 and ELF64} (Executable and Linkable Format) object files, as used by Linux as well as \i{Unix System V},
5668 including \i{Solaris x86}, \i{UnixWare} and \i{SCO Unix}. \c{elf}
5669 provides a default output file-name extension of \c{.o}.
5670 \c{elf} is a synonym for \c{elf32}.
5672 \S{abisect} ELF specific directive \i\c{osabi}
5674 The ELF header specifies the application binary interface for the target operating system (OSABI).
5675 This field can be set by using the \c{osabi} directive with the numeric value (0-255) of the target
5676 system. If this directive is not used, the default value will be "UNIX System V ABI" (0) which will work on
5677 most systems which support ELF.
5679 \S{elfsect} \c{elf} Extensions to the \c{SECTION}
5680 Directive\I{SECTION, elf extensions to}
5682 Like the \c{obj} format, \c{elf} allows you to specify additional
5683 information on the \c{SECTION} directive line, to control the type
5684 and properties of sections you declare. Section types and properties
5685 are generated automatically by NASM for the \i{standard section
5686 names}, but may still be
5687 overridden by these qualifiers.
5689 The available qualifiers are:
5691 \b \i\c{alloc} defines the section to be one which is loaded into
5692 memory when the program is run. \i\c{noalloc} defines it to be one
5693 which is not, such as an informational or comment section.
5695 \b \i\c{exec} defines the section to be one which should have execute
5696 permission when the program is run. \i\c{noexec} defines it as one
5699 \b \i\c{write} defines the section to be one which should be writable
5700 when the program is run. \i\c{nowrite} defines it as one which should
5703 \b \i\c{progbits} defines the section to be one with explicit contents
5704 stored in the object file: an ordinary code or data section, for
5705 example, \i\c{nobits} defines the section to be one with no explicit
5706 contents given, such as a BSS section.
5708 \b \c{align=}, used with a trailing number as in \c{obj}, gives the
5709 \I{section alignment, in elf}\I{alignment, in elf sections}alignment
5710 requirements of the section.
5712 \b \i\c{tls} defines the section to be one which contains
5713 thread local variables.
5715 The defaults assumed by NASM if you do not specify the above
5718 \I\c{.text} \I\c{.rodata} \I\c{.lrodata} \I\c{.data} \I\c{.ldata}
5719 \I\c{.bss} \I\c{.lbss} \I\c{.tdata} \I\c{.tbss} \I\c\{.comment}
5721 \c section .text progbits alloc exec nowrite align=16
5722 \c section .rodata progbits alloc noexec nowrite align=4
5723 \c section .lrodata progbits alloc noexec nowrite align=4
5724 \c section .data progbits alloc noexec write align=4
5725 \c section .ldata progbits alloc noexec write align=4
5726 \c section .bss nobits alloc noexec write align=4
5727 \c section .lbss nobits alloc noexec write align=4
5728 \c section .tdata progbits alloc noexec write align=4 tls
5729 \c section .tbss nobits alloc noexec write align=4 tls
5730 \c section .comment progbits noalloc noexec nowrite align=1
5731 \c section other progbits alloc noexec nowrite align=1
5733 (Any section name other than those in the above table
5734 is treated by default like \c{other} in the above table.
5735 Please note that section names are case sensitive.)
5738 \S{elfwrt} \i{Position-Independent Code}\I{PIC}: \c{elf} Special
5739 Symbols and \i\c{WRT}
5741 The \c{ELF} specification contains enough features to allow
5742 position-independent code (PIC) to be written, which makes \i{ELF
5743 shared libraries} very flexible. However, it also means NASM has to
5744 be able to generate a variety of ELF specific relocation types in ELF
5745 object files, if it is to be an assembler which can write PIC.
5747 Since \c{ELF} does not support segment-base references, the \c{WRT}
5748 operator is not used for its normal purpose; therefore NASM's
5749 \c{elf} output format makes use of \c{WRT} for a different purpose,
5750 namely the PIC-specific \I{relocations, PIC-specific}relocation
5753 \c{elf} defines five special symbols which you can use as the
5754 right-hand side of the \c{WRT} operator to obtain PIC relocation
5755 types. They are \i\c{..gotpc}, \i\c{..gotoff}, \i\c{..got},
5756 \i\c{..plt} and \i\c{..sym}. Their functions are summarized here:
5758 \b Referring to the symbol marking the global offset table base
5759 using \c{wrt ..gotpc} will end up giving the distance from the
5760 beginning of the current section to the global offset table.
5761 (\i\c{_GLOBAL_OFFSET_TABLE_} is the standard symbol name used to
5762 refer to the \i{GOT}.) So you would then need to add \i\c{$$} to the
5763 result to get the real address of the GOT.
5765 \b Referring to a location in one of your own sections using \c{wrt
5766 ..gotoff} will give the distance from the beginning of the GOT to
5767 the specified location, so that adding on the address of the GOT
5768 would give the real address of the location you wanted.
5770 \b Referring to an external or global symbol using \c{wrt ..got}
5771 causes the linker to build an entry \e{in} the GOT containing the
5772 address of the symbol, and the reference gives the distance from the
5773 beginning of the GOT to the entry; so you can add on the address of
5774 the GOT, load from the resulting address, and end up with the
5775 address of the symbol.
5777 \b Referring to a procedure name using \c{wrt ..plt} causes the
5778 linker to build a \i{procedure linkage table} entry for the symbol,
5779 and the reference gives the address of the \i{PLT} entry. You can
5780 only use this in contexts which would generate a PC-relative
5781 relocation normally (i.e. as the destination for \c{CALL} or
5782 \c{JMP}), since ELF contains no relocation type to refer to PLT
5785 \b Referring to a symbol name using \c{wrt ..sym} causes NASM to
5786 write an ordinary relocation, but instead of making the relocation
5787 relative to the start of the section and then adding on the offset
5788 to the symbol, it will write a relocation record aimed directly at
5789 the symbol in question. The distinction is a necessary one due to a
5790 peculiarity of the dynamic linker.
5792 A fuller explanation of how to use these relocation types to write
5793 shared libraries entirely in NASM is given in \k{picdll}.
5795 \S{elftls} \i{Thread Local Storage}\I{TLS}: \c{elf} Special
5796 Symbols and \i\c{WRT}
5798 \b In ELF32 mode, referring to an external or global symbol using
5799 \c{wrt ..tlsie} \I\c{..tlsie}
5800 causes the linker to build an entry \e{in} the GOT containing the
5801 offset of the symbol within the TLS block, so you can access the value
5802 of the symbol with code such as:
5804 \c mov eax,[tid wrt ..tlsie]
5808 \b In ELF64 mode, referring to an external or global symbol using
5809 \c{wrt ..gottpoff} \I\c{..gottpoff}
5810 causes the linker to build an entry \e{in} the GOT containing the
5811 offset of the symbol within the TLS block, so you can access the value
5812 of the symbol with code such as:
5814 \c mov rax,[rel tid wrt ..gottpoff]
5818 \S{elfglob} \c{elf} Extensions to the \c{GLOBAL} Directive\I{GLOBAL,
5819 elf extensions to}\I{GLOBAL, aoutb extensions to}
5821 \c{ELF} object files can contain more information about a global symbol
5822 than just its address: they can contain the \I{symbol sizes,
5823 specifying}\I{size, of symbols}size of the symbol and its \I{symbol
5824 types, specifying}\I{type, of symbols}type as well. These are not
5825 merely debugger conveniences, but are actually necessary when the
5826 program being written is a \i{shared library}. NASM therefore
5827 supports some extensions to the \c{GLOBAL} directive, allowing you
5828 to specify these features.
5830 You can specify whether a global variable is a function or a data
5831 object by suffixing the name with a colon and the word
5832 \i\c{function} or \i\c{data}. (\i\c{object} is a synonym for
5833 \c{data}.) For example:
5835 \c global hashlookup:function, hashtable:data
5837 exports the global symbol \c{hashlookup} as a function and
5838 \c{hashtable} as a data object.
5840 Optionally, you can control the ELF visibility of the symbol. Just
5841 add one of the visibility keywords: \i\c{default}, \i\c{internal},
5842 \i\c{hidden}, or \i\c{protected}. The default is \i\c{default} of
5843 course. For example, to make \c{hashlookup} hidden:
5845 \c global hashlookup:function hidden
5847 You can also specify the size of the data associated with the
5848 symbol, as a numeric expression (which may involve labels, and even
5849 forward references) after the type specifier. Like this:
5851 \c global hashtable:data (hashtable.end - hashtable)
5854 \c db this,that,theother ; some data here
5857 This makes NASM automatically calculate the length of the table and
5858 place that information into the \c{ELF} symbol table.
5860 Declaring the type and size of global symbols is necessary when
5861 writing shared library code. For more information, see
5865 \S{elfcomm} \c{elf} Extensions to the \c{COMMON} Directive
5866 \I{COMMON, elf extensions to}
5868 \c{ELF} also allows you to specify alignment requirements \I{common
5869 variables, alignment in elf}\I{alignment, of elf common variables}on
5870 common variables. This is done by putting a number (which must be a
5871 power of two) after the name and size of the common variable,
5872 separated (as usual) by a colon. For example, an array of
5873 doublewords would benefit from 4-byte alignment:
5875 \c common dwordarray 128:4
5877 This declares the total size of the array to be 128 bytes, and
5878 requires that it be aligned on a 4-byte boundary.
5881 \S{elf16} 16-bit code and ELF
5882 \I{ELF, 16-bit code and}
5884 The \c{ELF32} specification doesn't provide relocations for 8- and
5885 16-bit values, but the GNU \c{ld} linker adds these as an extension.
5886 NASM can generate GNU-compatible relocations, to allow 16-bit code to
5887 be linked as ELF using GNU \c{ld}. If NASM is used with the
5888 \c{-w+gnu-elf-extensions} option, a warning is issued when one of
5889 these relocations is generated.
5891 \S{elfdbg} Debug formats and ELF
5892 \I{ELF, Debug formats and}
5894 \c{ELF32} and \c{ELF64} provide debug information in \c{STABS} and \c{DWARF} formats.
5895 Line number information is generated for all executable sections, but please
5896 note that only the ".text" section is executable by default.
5898 \H{aoutfmt} \i\c{aout}: Linux \I{a.out, Linux version}\I{linux, a.out}\c{a.out} Object Files
5900 The \c{aout} format generates \c{a.out} object files, in the form used
5901 by early Linux systems (current Linux systems use ELF, see
5902 \k{elffmt}.) These differ from other \c{a.out} object files in that
5903 the magic number in the first four bytes of the file is
5904 different; also, some implementations of \c{a.out}, for example
5905 NetBSD's, support position-independent code, which Linux's
5906 implementation does not.
5908 \c{a.out} provides a default output file-name extension of \c{.o}.
5910 \c{a.out} is a very simple object format. It supports no special
5911 directives, no special symbols, no use of \c{SEG} or \c{WRT}, and no
5912 extensions to any standard directives. It supports only the three
5913 \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}.
5916 \H{aoutfmt} \i\c{aoutb}: \i{NetBSD}/\i{FreeBSD}/\i{OpenBSD}
5917 \I{a.out, BSD version}\c{a.out} Object Files
5919 The \c{aoutb} format generates \c{a.out} object files, in the form
5920 used by the various free \c{BSD Unix} clones, \c{NetBSD}, \c{FreeBSD}
5921 and \c{OpenBSD}. For simple object files, this object format is exactly
5922 the same as \c{aout} except for the magic number in the first four bytes
5923 of the file. However, the \c{aoutb} format supports
5924 \I{PIC}\i{position-independent code} in the same way as the \c{elf}
5925 format, so you can use it to write \c{BSD} \i{shared libraries}.
5927 \c{aoutb} provides a default output file-name extension of \c{.o}.
5929 \c{aoutb} supports no special directives, no special symbols, and
5930 only the three \i{standard section names} \i\c{.text}, \i\c{.data}
5931 and \i\c{.bss}. However, it also supports the same use of \i\c{WRT} as
5932 \c{elf} does, to provide position-independent code relocation types.
5933 See \k{elfwrt} for full documentation of this feature.
5935 \c{aoutb} also supports the same extensions to the \c{GLOBAL}
5936 directive as \c{elf} does: see \k{elfglob} for documentation of
5940 \H{as86fmt} \c{as86}: \i{Minix}/Linux\I{linux, as86} \i\c{as86} Object Files
5942 The Minix/Linux 16-bit assembler \c{as86} has its own non-standard
5943 object file format. Although its companion linker \i\c{ld86} produces
5944 something close to ordinary \c{a.out} binaries as output, the object
5945 file format used to communicate between \c{as86} and \c{ld86} is not
5948 NASM supports this format, just in case it is useful, as \c{as86}.
5949 \c{as86} provides a default output file-name extension of \c{.o}.
5951 \c{as86} is a very simple object format (from the NASM user's point
5952 of view). It supports no special directives, no use of \c{SEG} or \c{WRT},
5953 and no extensions to any standard directives. It supports only the three
5954 \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}. The
5955 only special symbol supported is \c{..start}.
5958 \H{rdffmt} \I{RDOFF}\i\c{rdf}: \i{Relocatable Dynamic Object File
5961 The \c{rdf} output format produces \c{RDOFF} object files. \c{RDOFF}
5962 (Relocatable Dynamic Object File Format) is a home-grown object-file
5963 format, designed alongside NASM itself and reflecting in its file
5964 format the internal structure of the assembler.
5966 \c{RDOFF} is not used by any well-known operating systems. Those
5967 writing their own systems, however, may well wish to use \c{RDOFF}
5968 as their object format, on the grounds that it is designed primarily
5969 for simplicity and contains very little file-header bureaucracy.
5971 The Unix NASM archive, and the DOS archive which includes sources,
5972 both contain an \I{rdoff subdirectory}\c{rdoff} subdirectory holding
5973 a set of RDOFF utilities: an RDF linker, an \c{RDF} static-library
5974 manager, an RDF file dump utility, and a program which will load and
5975 execute an RDF executable under Linux.
5977 \c{rdf} supports only the \i{standard section names} \i\c{.text},
5978 \i\c{.data} and \i\c{.bss}.
5981 \S{rdflib} Requiring a Library: The \i\c{LIBRARY} Directive
5983 \c{RDOFF} contains a mechanism for an object file to demand a given
5984 library to be linked to the module, either at load time or run time.
5985 This is done by the \c{LIBRARY} directive, which takes one argument
5986 which is the name of the module:
5988 \c library mylib.rdl
5991 \S{rdfmod} Specifying a Module Name: The \i\c{MODULE} Directive
5993 Special \c{RDOFF} header record is used to store the name of the module.
5994 It can be used, for example, by run-time loader to perform dynamic
5995 linking. \c{MODULE} directive takes one argument which is the name
6000 Note that when you statically link modules and tell linker to strip
6001 the symbols from output file, all module names will be stripped too.
6002 To avoid it, you should start module names with \I{$, prefix}\c{$}, like:
6004 \c module $kernel.core
6007 \S{rdfglob} \c{rdf} Extensions to the \c{GLOBAL} Directive\I{GLOBAL,
6010 \c{RDOFF} global symbols can contain additional information needed by
6011 the static linker. You can mark a global symbol as exported, thus
6012 telling the linker do not strip it from target executable or library
6013 file. Like in \c{ELF}, you can also specify whether an exported symbol
6014 is a procedure (function) or data object.
6016 Suffixing the name with a colon and the word \i\c{export} you make the
6019 \c global sys_open:export
6021 To specify that exported symbol is a procedure (function), you add the
6022 word \i\c{proc} or \i\c{function} after declaration:
6024 \c global sys_open:export proc
6026 Similarly, to specify exported data object, add the word \i\c{data}
6027 or \i\c{object} to the directive:
6029 \c global kernel_ticks:export data
6032 \S{rdfimpt} \c{rdf} Extensions to the \c{EXTERN} Directive\I{EXTERN,
6035 By default the \c{EXTERN} directive in \c{RDOFF} declares a "pure external"
6036 symbol (i.e. the static linker will complain if such a symbol is not resolved).
6037 To declare an "imported" symbol, which must be resolved later during a dynamic
6038 linking phase, \c{RDOFF} offers an additional \c{import} modifier. As in
6039 \c{GLOBAL}, you can also specify whether an imported symbol is a procedure
6040 (function) or data object. For example:
6043 \c extern _open:import
6044 \c extern _printf:import proc
6045 \c extern _errno:import data
6047 Here the directive \c{LIBRARY} is also included, which gives the dynamic linker
6048 a hint as to where to find requested symbols.
6051 \H{dbgfmt} \i\c{dbg}: Debugging Format
6053 The \c{dbg} output format is not built into NASM in the default
6054 configuration. If you are building your own NASM executable from the
6055 sources, you can define \i\c{OF_DBG} in \c{output/outform.h} or on the
6056 compiler command line, and obtain the \c{dbg} output format.
6058 The \c{dbg} format does not output an object file as such; instead,
6059 it outputs a text file which contains a complete list of all the
6060 transactions between the main body of NASM and the output-format
6061 back end module. It is primarily intended to aid people who want to
6062 write their own output drivers, so that they can get a clearer idea
6063 of the various requests the main program makes of the output driver,
6064 and in what order they happen.
6066 For simple files, one can easily use the \c{dbg} format like this:
6068 \c nasm -f dbg filename.asm
6070 which will generate a diagnostic file called \c{filename.dbg}.
6071 However, this will not work well on files which were designed for a
6072 different object format, because each object format defines its own
6073 macros (usually user-level forms of directives), and those macros
6074 will not be defined in the \c{dbg} format. Therefore it can be
6075 useful to run NASM twice, in order to do the preprocessing with the
6076 native object format selected:
6078 \c nasm -e -f rdf -o rdfprog.i rdfprog.asm
6079 \c nasm -a -f dbg rdfprog.i
6081 This preprocesses \c{rdfprog.asm} into \c{rdfprog.i}, keeping the
6082 \c{rdf} object format selected in order to make sure RDF special
6083 directives are converted into primitive form correctly. Then the
6084 preprocessed source is fed through the \c{dbg} format to generate
6085 the final diagnostic output.
6087 This workaround will still typically not work for programs intended
6088 for \c{obj} format, because the \c{obj} \c{SEGMENT} and \c{GROUP}
6089 directives have side effects of defining the segment and group names
6090 as symbols; \c{dbg} will not do this, so the program will not
6091 assemble. You will have to work around that by defining the symbols
6092 yourself (using \c{EXTERN}, for example) if you really need to get a
6093 \c{dbg} trace of an \c{obj}-specific source file.
6095 \c{dbg} accepts any section name and any directives at all, and logs
6096 them all to its output file.
6099 \C{16bit} Writing 16-bit Code (DOS, Windows 3/3.1)
6101 This chapter attempts to cover some of the common issues encountered
6102 when writing 16-bit code to run under \c{MS-DOS} or \c{Windows 3.x}. It
6103 covers how to link programs to produce \c{.EXE} or \c{.COM} files,
6104 how to write \c{.SYS} device drivers, and how to interface assembly
6105 language code with 16-bit C compilers and with Borland Pascal.
6108 \H{exefiles} Producing \i\c{.EXE} Files
6110 Any large program written under DOS needs to be built as a \c{.EXE}
6111 file: only \c{.EXE} files have the necessary internal structure
6112 required to span more than one 64K segment. \i{Windows} programs,
6113 also, have to be built as \c{.EXE} files, since Windows does not
6114 support the \c{.COM} format.
6116 In general, you generate \c{.EXE} files by using the \c{obj} output
6117 format to produce one or more \i\c{.OBJ} files, and then linking
6118 them together using a linker. However, NASM also supports the direct
6119 generation of simple DOS \c{.EXE} files using the \c{bin} output
6120 format (by using \c{DB} and \c{DW} to construct the \c{.EXE} file
6121 header), and a macro package is supplied to do this. Thanks to
6122 Yann Guidon for contributing the code for this.
6124 NASM may also support \c{.EXE} natively as another output format in
6128 \S{objexe} Using the \c{obj} Format To Generate \c{.EXE} Files
6130 This section describes the usual method of generating \c{.EXE} files
6131 by linking \c{.OBJ} files together.
6133 Most 16-bit programming language packages come with a suitable
6134 linker; if you have none of these, there is a free linker called
6135 \i{VAL}\I{linker, free}, available in \c{LZH} archive format from
6136 \W{ftp://x2ftp.oulu.fi/pub/msdos/programming/lang/}\i\c{x2ftp.oulu.fi}.
6137 An LZH archiver can be found at
6138 \W{ftp://ftp.simtel.net/pub/simtelnet/msdos/arcers}\i\c{ftp.simtel.net}.
6139 There is another `free' linker (though this one doesn't come with
6140 sources) called \i{FREELINK}, available from
6141 \W{http://www.pcorner.com/tpc/old/3-101.html}\i\c{www.pcorner.com}.
6142 A third, \i\c{djlink}, written by DJ Delorie, is available at
6143 \W{http://www.delorie.com/djgpp/16bit/djlink/}\i\c{www.delorie.com}.
6144 A fourth linker, \i\c{ALINK}, written by Anthony A.J. Williams, is
6145 available at \W{http://alink.sourceforge.net}\i\c{alink.sourceforge.net}.
6147 When linking several \c{.OBJ} files into a \c{.EXE} file, you should
6148 ensure that exactly one of them has a start point defined (using the
6149 \I{program entry point}\i\c{..start} special symbol defined by the
6150 \c{obj} format: see \k{dotdotstart}). If no module defines a start
6151 point, the linker will not know what value to give the entry-point
6152 field in the output file header; if more than one defines a start
6153 point, the linker will not know \e{which} value to use.
6155 An example of a NASM source file which can be assembled to a
6156 \c{.OBJ} file and linked on its own to a \c{.EXE} is given here. It
6157 demonstrates the basic principles of defining a stack, initialising
6158 the segment registers, and declaring a start point. This file is
6159 also provided in the \I{test subdirectory}\c{test} subdirectory of
6160 the NASM archives, under the name \c{objexe.asm}.
6171 This initial piece of code sets up \c{DS} to point to the data
6172 segment, and initializes \c{SS} and \c{SP} to point to the top of
6173 the provided stack. Notice that interrupts are implicitly disabled
6174 for one instruction after a move into \c{SS}, precisely for this
6175 situation, so that there's no chance of an interrupt occurring
6176 between the loads of \c{SS} and \c{SP} and not having a stack to
6179 Note also that the special symbol \c{..start} is defined at the
6180 beginning of this code, which means that will be the entry point
6181 into the resulting executable file.
6187 The above is the main program: load \c{DS:DX} with a pointer to the
6188 greeting message (\c{hello} is implicitly relative to the segment
6189 \c{data}, which was loaded into \c{DS} in the setup code, so the
6190 full pointer is valid), and call the DOS print-string function.
6195 This terminates the program using another DOS system call.
6199 \c hello: db 'hello, world', 13, 10, '$'
6201 The data segment contains the string we want to display.
6203 \c segment stack stack
6207 The above code declares a stack segment containing 64 bytes of
6208 uninitialized stack space, and points \c{stacktop} at the top of it.
6209 The directive \c{segment stack stack} defines a segment \e{called}
6210 \c{stack}, and also of \e{type} \c{STACK}. The latter is not
6211 necessary to the correct running of the program, but linkers are
6212 likely to issue warnings or errors if your program has no segment of
6215 The above file, when assembled into a \c{.OBJ} file, will link on
6216 its own to a valid \c{.EXE} file, which when run will print `hello,
6217 world' and then exit.
6220 \S{binexe} Using the \c{bin} Format To Generate \c{.EXE} Files
6222 The \c{.EXE} file format is simple enough that it's possible to
6223 build a \c{.EXE} file by writing a pure-binary program and sticking
6224 a 32-byte header on the front. This header is simple enough that it
6225 can be generated using \c{DB} and \c{DW} commands by NASM itself, so
6226 that you can use the \c{bin} output format to directly generate
6229 Included in the NASM archives, in the \I{misc subdirectory}\c{misc}
6230 subdirectory, is a file \i\c{exebin.mac} of macros. It defines three
6231 macros: \i\c{EXE_begin}, \i\c{EXE_stack} and \i\c{EXE_end}.
6233 To produce a \c{.EXE} file using this method, you should start by
6234 using \c{%include} to load the \c{exebin.mac} macro package into
6235 your source file. You should then issue the \c{EXE_begin} macro call
6236 (which takes no arguments) to generate the file header data. Then
6237 write code as normal for the \c{bin} format - you can use all three
6238 standard sections \c{.text}, \c{.data} and \c{.bss}. At the end of
6239 the file you should call the \c{EXE_end} macro (again, no arguments),
6240 which defines some symbols to mark section sizes, and these symbols
6241 are referred to in the header code generated by \c{EXE_begin}.
6243 In this model, the code you end up writing starts at \c{0x100}, just
6244 like a \c{.COM} file - in fact, if you strip off the 32-byte header
6245 from the resulting \c{.EXE} file, you will have a valid \c{.COM}
6246 program. All the segment bases are the same, so you are limited to a
6247 64K program, again just like a \c{.COM} file. Note that an \c{ORG}
6248 directive is issued by the \c{EXE_begin} macro, so you should not
6249 explicitly issue one of your own.
6251 You can't directly refer to your segment base value, unfortunately,
6252 since this would require a relocation in the header, and things
6253 would get a lot more complicated. So you should get your segment
6254 base by copying it out of \c{CS} instead.
6256 On entry to your \c{.EXE} file, \c{SS:SP} are already set up to
6257 point to the top of a 2Kb stack. You can adjust the default stack
6258 size of 2Kb by calling the \c{EXE_stack} macro. For example, to
6259 change the stack size of your program to 64 bytes, you would call
6262 A sample program which generates a \c{.EXE} file in this way is
6263 given in the \c{test} subdirectory of the NASM archive, as
6267 \H{comfiles} Producing \i\c{.COM} Files
6269 While large DOS programs must be written as \c{.EXE} files, small
6270 ones are often better written as \c{.COM} files. \c{.COM} files are
6271 pure binary, and therefore most easily produced using the \c{bin}
6275 \S{combinfmt} Using the \c{bin} Format To Generate \c{.COM} Files
6277 \c{.COM} files expect to be loaded at offset \c{100h} into their
6278 segment (though the segment may change). Execution then begins at
6279 \I\c{ORG}\c{100h}, i.e. right at the start of the program. So to
6280 write a \c{.COM} program, you would create a source file looking
6288 \c ; put your code here
6292 \c ; put data items here
6296 \c ; put uninitialized data here
6298 The \c{bin} format puts the \c{.text} section first in the file, so
6299 you can declare data or BSS items before beginning to write code if
6300 you want to and the code will still end up at the front of the file
6303 The BSS (uninitialized data) section does not take up space in the
6304 \c{.COM} file itself: instead, addresses of BSS items are resolved
6305 to point at space beyond the end of the file, on the grounds that
6306 this will be free memory when the program is run. Therefore you
6307 should not rely on your BSS being initialized to all zeros when you
6310 To assemble the above program, you should use a command line like
6312 \c nasm myprog.asm -fbin -o myprog.com
6314 The \c{bin} format would produce a file called \c{myprog} if no
6315 explicit output file name were specified, so you have to override it
6316 and give the desired file name.
6319 \S{comobjfmt} Using the \c{obj} Format To Generate \c{.COM} Files
6321 If you are writing a \c{.COM} program as more than one module, you
6322 may wish to assemble several \c{.OBJ} files and link them together
6323 into a \c{.COM} program. You can do this, provided you have a linker
6324 capable of outputting \c{.COM} files directly (\i{TLINK} does this),
6325 or alternatively a converter program such as \i\c{EXE2BIN} to
6326 transform the \c{.EXE} file output from the linker into a \c{.COM}
6329 If you do this, you need to take care of several things:
6331 \b The first object file containing code should start its code
6332 segment with a line like \c{RESB 100h}. This is to ensure that the
6333 code begins at offset \c{100h} relative to the beginning of the code
6334 segment, so that the linker or converter program does not have to
6335 adjust address references within the file when generating the
6336 \c{.COM} file. Other assemblers use an \i\c{ORG} directive for this
6337 purpose, but \c{ORG} in NASM is a format-specific directive to the
6338 \c{bin} output format, and does not mean the same thing as it does
6339 in MASM-compatible assemblers.
6341 \b You don't need to define a stack segment.
6343 \b All your segments should be in the same group, so that every time
6344 your code or data references a symbol offset, all offsets are
6345 relative to the same segment base. This is because, when a \c{.COM}
6346 file is loaded, all the segment registers contain the same value.
6349 \H{sysfiles} Producing \i\c{.SYS} Files
6351 \i{MS-DOS device drivers} - \c{.SYS} files - are pure binary files,
6352 similar to \c{.COM} files, except that they start at origin zero
6353 rather than \c{100h}. Therefore, if you are writing a device driver
6354 using the \c{bin} format, you do not need the \c{ORG} directive,
6355 since the default origin for \c{bin} is zero. Similarly, if you are
6356 using \c{obj}, you do not need the \c{RESB 100h} at the start of
6359 \c{.SYS} files start with a header structure, containing pointers to
6360 the various routines inside the driver which do the work. This
6361 structure should be defined at the start of the code segment, even
6362 though it is not actually code.
6364 For more information on the format of \c{.SYS} files, and the data
6365 which has to go in the header structure, a list of books is given in
6366 the Frequently Asked Questions list for the newsgroup
6367 \W{news:comp.os.msdos.programmer}\i\c{comp.os.msdos.programmer}.
6370 \H{16c} Interfacing to 16-bit C Programs
6372 This section covers the basics of writing assembly routines that
6373 call, or are called from, C programs. To do this, you would
6374 typically write an assembly module as a \c{.OBJ} file, and link it
6375 with your C modules to produce a \i{mixed-language program}.
6378 \S{16cunder} External Symbol Names
6380 \I{C symbol names}\I{underscore, in C symbols}C compilers have the
6381 convention that the names of all global symbols (functions or data)
6382 they define are formed by prefixing an underscore to the name as it
6383 appears in the C program. So, for example, the function a C
6384 programmer thinks of as \c{printf} appears to an assembly language
6385 programmer as \c{_printf}. This means that in your assembly
6386 programs, you can define symbols without a leading underscore, and
6387 not have to worry about name clashes with C symbols.
6389 If you find the underscores inconvenient, you can define macros to
6390 replace the \c{GLOBAL} and \c{EXTERN} directives as follows:
6406 (These forms of the macros only take one argument at a time; a
6407 \c{%rep} construct could solve this.)
6409 If you then declare an external like this:
6413 then the macro will expand it as
6416 \c %define printf _printf
6418 Thereafter, you can reference \c{printf} as if it was a symbol, and
6419 the preprocessor will put the leading underscore on where necessary.
6421 The \c{cglobal} macro works similarly. You must use \c{cglobal}
6422 before defining the symbol in question, but you would have had to do
6423 that anyway if you used \c{GLOBAL}.
6425 Also see \k{opt-pfix}.
6427 \S{16cmodels} \i{Memory Models}
6429 NASM contains no mechanism to support the various C memory models
6430 directly; you have to keep track yourself of which one you are
6431 writing for. This means you have to keep track of the following
6434 \b In models using a single code segment (tiny, small and compact),
6435 functions are near. This means that function pointers, when stored
6436 in data segments or pushed on the stack as function arguments, are
6437 16 bits long and contain only an offset field (the \c{CS} register
6438 never changes its value, and always gives the segment part of the
6439 full function address), and that functions are called using ordinary
6440 near \c{CALL} instructions and return using \c{RETN} (which, in
6441 NASM, is synonymous with \c{RET} anyway). This means both that you
6442 should write your own routines to return with \c{RETN}, and that you
6443 should call external C routines with near \c{CALL} instructions.
6445 \b In models using more than one code segment (medium, large and
6446 huge), functions are far. This means that function pointers are 32
6447 bits long (consisting of a 16-bit offset followed by a 16-bit
6448 segment), and that functions are called using \c{CALL FAR} (or
6449 \c{CALL seg:offset}) and return using \c{RETF}. Again, you should
6450 therefore write your own routines to return with \c{RETF} and use
6451 \c{CALL FAR} to call external routines.
6453 \b In models using a single data segment (tiny, small and medium),
6454 data pointers are 16 bits long, containing only an offset field (the
6455 \c{DS} register doesn't change its value, and always gives the
6456 segment part of the full data item address).
6458 \b In models using more than one data segment (compact, large and
6459 huge), data pointers are 32 bits long, consisting of a 16-bit offset
6460 followed by a 16-bit segment. You should still be careful not to
6461 modify \c{DS} in your routines without restoring it afterwards, but
6462 \c{ES} is free for you to use to access the contents of 32-bit data
6463 pointers you are passed.
6465 \b The huge memory model allows single data items to exceed 64K in
6466 size. In all other memory models, you can access the whole of a data
6467 item just by doing arithmetic on the offset field of the pointer you
6468 are given, whether a segment field is present or not; in huge model,
6469 you have to be more careful of your pointer arithmetic.
6471 \b In most memory models, there is a \e{default} data segment, whose
6472 segment address is kept in \c{DS} throughout the program. This data
6473 segment is typically the same segment as the stack, kept in \c{SS},
6474 so that functions' local variables (which are stored on the stack)
6475 and global data items can both be accessed easily without changing
6476 \c{DS}. Particularly large data items are typically stored in other
6477 segments. However, some memory models (though not the standard
6478 ones, usually) allow the assumption that \c{SS} and \c{DS} hold the
6479 same value to be removed. Be careful about functions' local
6480 variables in this latter case.
6482 In models with a single code segment, the segment is called
6483 \i\c{_TEXT}, so your code segment must also go by this name in order
6484 to be linked into the same place as the main code segment. In models
6485 with a single data segment, or with a default data segment, it is
6489 \S{16cfunc} Function Definitions and Function Calls
6491 \I{functions, C calling convention}The \i{C calling convention} in
6492 16-bit programs is as follows. In the following description, the
6493 words \e{caller} and \e{callee} are used to denote the function
6494 doing the calling and the function which gets called.
6496 \b The caller pushes the function's parameters on the stack, one
6497 after another, in reverse order (right to left, so that the first
6498 argument specified to the function is pushed last).
6500 \b The caller then executes a \c{CALL} instruction to pass control
6501 to the callee. This \c{CALL} is either near or far depending on the
6504 \b The callee receives control, and typically (although this is not
6505 actually necessary, in functions which do not need to access their
6506 parameters) starts by saving the value of \c{SP} in \c{BP} so as to
6507 be able to use \c{BP} as a base pointer to find its parameters on
6508 the stack. However, the caller was probably doing this too, so part
6509 of the calling convention states that \c{BP} must be preserved by
6510 any C function. Hence the callee, if it is going to set up \c{BP} as
6511 a \i\e{frame pointer}, must push the previous value first.
6513 \b The callee may then access its parameters relative to \c{BP}.
6514 The word at \c{[BP]} holds the previous value of \c{BP} as it was
6515 pushed; the next word, at \c{[BP+2]}, holds the offset part of the
6516 return address, pushed implicitly by \c{CALL}. In a small-model
6517 (near) function, the parameters start after that, at \c{[BP+4]}; in
6518 a large-model (far) function, the segment part of the return address
6519 lives at \c{[BP+4]}, and the parameters begin at \c{[BP+6]}. The
6520 leftmost parameter of the function, since it was pushed last, is
6521 accessible at this offset from \c{BP}; the others follow, at
6522 successively greater offsets. Thus, in a function such as \c{printf}
6523 which takes a variable number of parameters, the pushing of the
6524 parameters in reverse order means that the function knows where to
6525 find its first parameter, which tells it the number and type of the
6528 \b The callee may also wish to decrease \c{SP} further, so as to
6529 allocate space on the stack for local variables, which will then be
6530 accessible at negative offsets from \c{BP}.
6532 \b The callee, if it wishes to return a value to the caller, should
6533 leave the value in \c{AL}, \c{AX} or \c{DX:AX} depending on the size
6534 of the value. Floating-point results are sometimes (depending on the
6535 compiler) returned in \c{ST0}.
6537 \b Once the callee has finished processing, it restores \c{SP} from
6538 \c{BP} if it had allocated local stack space, then pops the previous
6539 value of \c{BP}, and returns via \c{RETN} or \c{RETF} depending on
6542 \b When the caller regains control from the callee, the function
6543 parameters are still on the stack, so it typically adds an immediate
6544 constant to \c{SP} to remove them (instead of executing a number of
6545 slow \c{POP} instructions). Thus, if a function is accidentally
6546 called with the wrong number of parameters due to a prototype
6547 mismatch, the stack will still be returned to a sensible state since
6548 the caller, which \e{knows} how many parameters it pushed, does the
6551 It is instructive to compare this calling convention with that for
6552 Pascal programs (described in \k{16bpfunc}). Pascal has a simpler
6553 convention, since no functions have variable numbers of parameters.
6554 Therefore the callee knows how many parameters it should have been
6555 passed, and is able to deallocate them from the stack itself by
6556 passing an immediate argument to the \c{RET} or \c{RETF}
6557 instruction, so the caller does not have to do it. Also, the
6558 parameters are pushed in left-to-right order, not right-to-left,
6559 which means that a compiler can give better guarantees about
6560 sequence points without performance suffering.
6562 Thus, you would define a function in C style in the following way.
6563 The following example is for small model:
6570 \c sub sp,0x40 ; 64 bytes of local stack space
6571 \c mov bx,[bp+4] ; first parameter to function
6575 \c mov sp,bp ; undo "sub sp,0x40" above
6579 For a large-model function, you would replace \c{RET} by \c{RETF},
6580 and look for the first parameter at \c{[BP+6]} instead of
6581 \c{[BP+4]}. Of course, if one of the parameters is a pointer, then
6582 the offsets of \e{subsequent} parameters will change depending on
6583 the memory model as well: far pointers take up four bytes on the
6584 stack when passed as a parameter, whereas near pointers take up two.
6586 At the other end of the process, to call a C function from your
6587 assembly code, you would do something like this:
6591 \c ; and then, further down...
6593 \c push word [myint] ; one of my integer variables
6594 \c push word mystring ; pointer into my data segment
6596 \c add sp,byte 4 ; `byte' saves space
6598 \c ; then those data items...
6603 \c mystring db 'This number -> %d <- should be 1234',10,0
6605 This piece of code is the small-model assembly equivalent of the C
6608 \c int myint = 1234;
6609 \c printf("This number -> %d <- should be 1234\n", myint);
6611 In large model, the function-call code might look more like this. In
6612 this example, it is assumed that \c{DS} already holds the segment
6613 base of the segment \c{_DATA}. If not, you would have to initialize
6616 \c push word [myint]
6617 \c push word seg mystring ; Now push the segment, and...
6618 \c push word mystring ; ... offset of "mystring"
6622 The integer value still takes up one word on the stack, since large
6623 model does not affect the size of the \c{int} data type. The first
6624 argument (pushed last) to \c{printf}, however, is a data pointer,
6625 and therefore has to contain a segment and offset part. The segment
6626 should be stored second in memory, and therefore must be pushed
6627 first. (Of course, \c{PUSH DS} would have been a shorter instruction
6628 than \c{PUSH WORD SEG mystring}, if \c{DS} was set up as the above
6629 example assumed.) Then the actual call becomes a far call, since
6630 functions expect far calls in large model; and \c{SP} has to be
6631 increased by 6 rather than 4 afterwards to make up for the extra
6635 \S{16cdata} Accessing Data Items
6637 To get at the contents of C variables, or to declare variables which
6638 C can access, you need only declare the names as \c{GLOBAL} or
6639 \c{EXTERN}. (Again, the names require leading underscores, as stated
6640 in \k{16cunder}.) Thus, a C variable declared as \c{int i} can be
6641 accessed from assembler as
6647 And to declare your own integer variable which C programs can access
6648 as \c{extern int j}, you do this (making sure you are assembling in
6649 the \c{_DATA} segment, if necessary):
6655 To access a C array, you need to know the size of the components of
6656 the array. For example, \c{int} variables are two bytes long, so if
6657 a C program declares an array as \c{int a[10]}, you can access
6658 \c{a[3]} by coding \c{mov ax,[_a+6]}. (The byte offset 6 is obtained
6659 by multiplying the desired array index, 3, by the size of the array
6660 element, 2.) The sizes of the C base types in 16-bit compilers are:
6661 1 for \c{char}, 2 for \c{short} and \c{int}, 4 for \c{long} and
6662 \c{float}, and 8 for \c{double}.
6664 To access a C \i{data structure}, you need to know the offset from
6665 the base of the structure to the field you are interested in. You
6666 can either do this by converting the C structure definition into a
6667 NASM structure definition (using \i\c{STRUC}), or by calculating the
6668 one offset and using just that.
6670 To do either of these, you should read your C compiler's manual to
6671 find out how it organizes data structures. NASM gives no special
6672 alignment to structure members in its own \c{STRUC} macro, so you
6673 have to specify alignment yourself if the C compiler generates it.
6674 Typically, you might find that a structure like
6681 might be four bytes long rather than three, since the \c{int} field
6682 would be aligned to a two-byte boundary. However, this sort of
6683 feature tends to be a configurable option in the C compiler, either
6684 using command-line options or \c{#pragma} lines, so you have to find
6685 out how your own compiler does it.
6688 \S{16cmacro} \i\c{c16.mac}: Helper Macros for the 16-bit C Interface
6690 Included in the NASM archives, in the \I{misc subdirectory}\c{misc}
6691 directory, is a file \c{c16.mac} of macros. It defines three macros:
6692 \i\c{proc}, \i\c{arg} and \i\c{endproc}. These are intended to be
6693 used for C-style procedure definitions, and they automate a lot of
6694 the work involved in keeping track of the calling convention.
6696 (An alternative, TASM compatible form of \c{arg} is also now built
6697 into NASM's preprocessor. See \k{stackrel} for details.)
6699 An example of an assembly function using the macro set is given
6706 \c mov ax,[bp + %$i]
6707 \c mov bx,[bp + %$j]
6712 This defines \c{_nearproc} to be a procedure taking two arguments,
6713 the first (\c{i}) an integer and the second (\c{j}) a pointer to an
6714 integer. It returns \c{i + *j}.
6716 Note that the \c{arg} macro has an \c{EQU} as the first line of its
6717 expansion, and since the label before the macro call gets prepended
6718 to the first line of the expanded macro, the \c{EQU} works, defining
6719 \c{%$i} to be an offset from \c{BP}. A context-local variable is
6720 used, local to the context pushed by the \c{proc} macro and popped
6721 by the \c{endproc} macro, so that the same argument name can be used
6722 in later procedures. Of course, you don't \e{have} to do that.
6724 The macro set produces code for near functions (tiny, small and
6725 compact-model code) by default. You can have it generate far
6726 functions (medium, large and huge-model code) by means of coding
6727 \I\c{FARCODE}\c{%define FARCODE}. This changes the kind of return
6728 instruction generated by \c{endproc}, and also changes the starting
6729 point for the argument offsets. The macro set contains no intrinsic
6730 dependency on whether data pointers are far or not.
6732 \c{arg} can take an optional parameter, giving the size of the
6733 argument. If no size is given, 2 is assumed, since it is likely that
6734 many function parameters will be of type \c{int}.
6736 The large-model equivalent of the above function would look like this:
6744 \c mov ax,[bp + %$i]
6745 \c mov bx,[bp + %$j]
6746 \c mov es,[bp + %$j + 2]
6751 This makes use of the argument to the \c{arg} macro to define a
6752 parameter of size 4, because \c{j} is now a far pointer. When we
6753 load from \c{j}, we must load a segment and an offset.
6756 \H{16bp} Interfacing to \i{Borland Pascal} Programs
6758 Interfacing to Borland Pascal programs is similar in concept to
6759 interfacing to 16-bit C programs. The differences are:
6761 \b The leading underscore required for interfacing to C programs is
6762 not required for Pascal.
6764 \b The memory model is always large: functions are far, data
6765 pointers are far, and no data item can be more than 64K long.
6766 (Actually, some functions are near, but only those functions that
6767 are local to a Pascal unit and never called from outside it. All
6768 assembly functions that Pascal calls, and all Pascal functions that
6769 assembly routines are able to call, are far.) However, all static
6770 data declared in a Pascal program goes into the default data
6771 segment, which is the one whose segment address will be in \c{DS}
6772 when control is passed to your assembly code. The only things that
6773 do not live in the default data segment are local variables (they
6774 live in the stack segment) and dynamically allocated variables. All
6775 data \e{pointers}, however, are far.
6777 \b The function calling convention is different - described below.
6779 \b Some data types, such as strings, are stored differently.
6781 \b There are restrictions on the segment names you are allowed to
6782 use - Borland Pascal will ignore code or data declared in a segment
6783 it doesn't like the name of. The restrictions are described below.
6786 \S{16bpfunc} The Pascal Calling Convention
6788 \I{functions, Pascal calling convention}\I{Pascal calling
6789 convention}The 16-bit Pascal calling convention is as follows. In
6790 the following description, the words \e{caller} and \e{callee} are
6791 used to denote the function doing the calling and the function which
6794 \b The caller pushes the function's parameters on the stack, one
6795 after another, in normal order (left to right, so that the first
6796 argument specified to the function is pushed first).
6798 \b The caller then executes a far \c{CALL} instruction to pass
6799 control to the callee.
6801 \b The callee receives control, and typically (although this is not
6802 actually necessary, in functions which do not need to access their
6803 parameters) starts by saving the value of \c{SP} in \c{BP} so as to
6804 be able to use \c{BP} as a base pointer to find its parameters on
6805 the stack. However, the caller was probably doing this too, so part
6806 of the calling convention states that \c{BP} must be preserved by
6807 any function. Hence the callee, if it is going to set up \c{BP} as a
6808 \i{frame pointer}, must push the previous value first.
6810 \b The callee may then access its parameters relative to \c{BP}.
6811 The word at \c{[BP]} holds the previous value of \c{BP} as it was
6812 pushed. The next word, at \c{[BP+2]}, holds the offset part of the
6813 return address, and the next one at \c{[BP+4]} the segment part. The
6814 parameters begin at \c{[BP+6]}. The rightmost parameter of the
6815 function, since it was pushed last, is accessible at this offset
6816 from \c{BP}; the others follow, at successively greater offsets.
6818 \b The callee may also wish to decrease \c{SP} further, so as to
6819 allocate space on the stack for local variables, which will then be
6820 accessible at negative offsets from \c{BP}.
6822 \b The callee, if it wishes to return a value to the caller, should
6823 leave the value in \c{AL}, \c{AX} or \c{DX:AX} depending on the size
6824 of the value. Floating-point results are returned in \c{ST0}.
6825 Results of type \c{Real} (Borland's own custom floating-point data
6826 type, not handled directly by the FPU) are returned in \c{DX:BX:AX}.
6827 To return a result of type \c{String}, the caller pushes a pointer
6828 to a temporary string before pushing the parameters, and the callee
6829 places the returned string value at that location. The pointer is
6830 not a parameter, and should not be removed from the stack by the
6831 \c{RETF} instruction.
6833 \b Once the callee has finished processing, it restores \c{SP} from
6834 \c{BP} if it had allocated local stack space, then pops the previous
6835 value of \c{BP}, and returns via \c{RETF}. It uses the form of
6836 \c{RETF} with an immediate parameter, giving the number of bytes
6837 taken up by the parameters on the stack. This causes the parameters
6838 to be removed from the stack as a side effect of the return
6841 \b When the caller regains control from the callee, the function
6842 parameters have already been removed from the stack, so it needs to
6845 Thus, you would define a function in Pascal style, taking two
6846 \c{Integer}-type parameters, in the following way:
6852 \c sub sp,0x40 ; 64 bytes of local stack space
6853 \c mov bx,[bp+8] ; first parameter to function
6854 \c mov bx,[bp+6] ; second parameter to function
6858 \c mov sp,bp ; undo "sub sp,0x40" above
6860 \c retf 4 ; total size of params is 4
6862 At the other end of the process, to call a Pascal function from your
6863 assembly code, you would do something like this:
6867 \c ; and then, further down...
6869 \c push word seg mystring ; Now push the segment, and...
6870 \c push word mystring ; ... offset of "mystring"
6871 \c push word [myint] ; one of my variables
6872 \c call far SomeFunc
6874 This is equivalent to the Pascal code
6876 \c procedure SomeFunc(String: PChar; Int: Integer);
6877 \c SomeFunc(@mystring, myint);
6880 \S{16bpseg} Borland Pascal \I{segment names, Borland Pascal}Segment
6883 Since Borland Pascal's internal unit file format is completely
6884 different from \c{OBJ}, it only makes a very sketchy job of actually
6885 reading and understanding the various information contained in a
6886 real \c{OBJ} file when it links that in. Therefore an object file
6887 intended to be linked to a Pascal program must obey a number of
6890 \b Procedures and functions must be in a segment whose name is
6891 either \c{CODE}, \c{CSEG}, or something ending in \c{_TEXT}.
6893 \b initialized data must be in a segment whose name is either
6894 \c{CONST} or something ending in \c{_DATA}.
6896 \b Uninitialized data must be in a segment whose name is either
6897 \c{DATA}, \c{DSEG}, or something ending in \c{_BSS}.
6899 \b Any other segments in the object file are completely ignored.
6900 \c{GROUP} directives and segment attributes are also ignored.
6903 \S{16bpmacro} Using \i\c{c16.mac} With Pascal Programs
6905 The \c{c16.mac} macro package, described in \k{16cmacro}, can also
6906 be used to simplify writing functions to be called from Pascal
6907 programs, if you code \I\c{PASCAL}\c{%define PASCAL}. This
6908 definition ensures that functions are far (it implies
6909 \i\c{FARCODE}), and also causes procedure return instructions to be
6910 generated with an operand.
6912 Defining \c{PASCAL} does not change the code which calculates the
6913 argument offsets; you must declare your function's arguments in
6914 reverse order. For example:
6922 \c mov ax,[bp + %$i]
6923 \c mov bx,[bp + %$j]
6924 \c mov es,[bp + %$j + 2]
6929 This defines the same routine, conceptually, as the example in
6930 \k{16cmacro}: it defines a function taking two arguments, an integer
6931 and a pointer to an integer, which returns the sum of the integer
6932 and the contents of the pointer. The only difference between this
6933 code and the large-model C version is that \c{PASCAL} is defined
6934 instead of \c{FARCODE}, and that the arguments are declared in
6938 \C{32bit} Writing 32-bit Code (Unix, Win32, DJGPP)
6940 This chapter attempts to cover some of the common issues involved
6941 when writing 32-bit code, to run under \i{Win32} or Unix, or to be
6942 linked with C code generated by a Unix-style C compiler such as
6943 \i{DJGPP}. It covers how to write assembly code to interface with
6944 32-bit C routines, and how to write position-independent code for
6947 Almost all 32-bit code, and in particular all code running under
6948 \c{Win32}, \c{DJGPP} or any of the PC Unix variants, runs in \I{flat
6949 memory model}\e{flat} memory model. This means that the segment registers
6950 and paging have already been set up to give you the same 32-bit 4Gb
6951 address space no matter what segment you work relative to, and that
6952 you should ignore all segment registers completely. When writing
6953 flat-model application code, you never need to use a segment
6954 override or modify any segment register, and the code-section
6955 addresses you pass to \c{CALL} and \c{JMP} live in the same address
6956 space as the data-section addresses you access your variables by and
6957 the stack-section addresses you access local variables and procedure
6958 parameters by. Every address is 32 bits long and contains only an
6962 \H{32c} Interfacing to 32-bit C Programs
6964 A lot of the discussion in \k{16c}, about interfacing to 16-bit C
6965 programs, still applies when working in 32 bits. The absence of
6966 memory models or segmentation worries simplifies things a lot.
6969 \S{32cunder} External Symbol Names
6971 Most 32-bit C compilers share the convention used by 16-bit
6972 compilers, that the names of all global symbols (functions or data)
6973 they define are formed by prefixing an underscore to the name as it
6974 appears in the C program. However, not all of them do: the \c{ELF}
6975 specification states that C symbols do \e{not} have a leading
6976 underscore on their assembly-language names.
6978 The older Linux \c{a.out} C compiler, all \c{Win32} compilers,
6979 \c{DJGPP}, and \c{NetBSD} and \c{FreeBSD}, all use the leading
6980 underscore; for these compilers, the macros \c{cextern} and
6981 \c{cglobal}, as given in \k{16cunder}, will still work. For \c{ELF},
6982 though, the leading underscore should not be used.
6984 See also \k{opt-pfix}.
6986 \S{32cfunc} Function Definitions and Function Calls
6988 \I{functions, C calling convention}The \i{C calling convention}
6989 in 32-bit programs is as follows. In the following description,
6990 the words \e{caller} and \e{callee} are used to denote
6991 the function doing the calling and the function which gets called.
6993 \b The caller pushes the function's parameters on the stack, one
6994 after another, in reverse order (right to left, so that the first
6995 argument specified to the function is pushed last).
6997 \b The caller then executes a near \c{CALL} instruction to pass
6998 control to the callee.
7000 \b The callee receives control, and typically (although this is not
7001 actually necessary, in functions which do not need to access their
7002 parameters) starts by saving the value of \c{ESP} in \c{EBP} so as
7003 to be able to use \c{EBP} as a base pointer to find its parameters
7004 on the stack. However, the caller was probably doing this too, so
7005 part of the calling convention states that \c{EBP} must be preserved
7006 by any C function. Hence the callee, if it is going to set up
7007 \c{EBP} as a \i{frame pointer}, must push the previous value first.
7009 \b The callee may then access its parameters relative to \c{EBP}.
7010 The doubleword at \c{[EBP]} holds the previous value of \c{EBP} as
7011 it was pushed; the next doubleword, at \c{[EBP+4]}, holds the return
7012 address, pushed implicitly by \c{CALL}. The parameters start after
7013 that, at \c{[EBP+8]}. The leftmost parameter of the function, since
7014 it was pushed last, is accessible at this offset from \c{EBP}; the
7015 others follow, at successively greater offsets. Thus, in a function
7016 such as \c{printf} which takes a variable number of parameters, the
7017 pushing of the parameters in reverse order means that the function
7018 knows where to find its first parameter, which tells it the number
7019 and type of the remaining ones.
7021 \b The callee may also wish to decrease \c{ESP} further, so as to
7022 allocate space on the stack for local variables, which will then be
7023 accessible at negative offsets from \c{EBP}.
7025 \b The callee, if it wishes to return a value to the caller, should
7026 leave the value in \c{AL}, \c{AX} or \c{EAX} depending on the size
7027 of the value. Floating-point results are typically returned in
7030 \b Once the callee has finished processing, it restores \c{ESP} from
7031 \c{EBP} if it had allocated local stack space, then pops the previous
7032 value of \c{EBP}, and returns via \c{RET} (equivalently, \c{RETN}).
7034 \b When the caller regains control from the callee, the function
7035 parameters are still on the stack, so it typically adds an immediate
7036 constant to \c{ESP} to remove them (instead of executing a number of
7037 slow \c{POP} instructions). Thus, if a function is accidentally
7038 called with the wrong number of parameters due to a prototype
7039 mismatch, the stack will still be returned to a sensible state since
7040 the caller, which \e{knows} how many parameters it pushed, does the
7043 There is an alternative calling convention used by Win32 programs
7044 for Windows API calls, and also for functions called \e{by} the
7045 Windows API such as window procedures: they follow what Microsoft
7046 calls the \c{__stdcall} convention. This is slightly closer to the
7047 Pascal convention, in that the callee clears the stack by passing a
7048 parameter to the \c{RET} instruction. However, the parameters are
7049 still pushed in right-to-left order.
7051 Thus, you would define a function in C style in the following way:
7058 \c sub esp,0x40 ; 64 bytes of local stack space
7059 \c mov ebx,[ebp+8] ; first parameter to function
7063 \c leave ; mov esp,ebp / pop ebp
7066 At the other end of the process, to call a C function from your
7067 assembly code, you would do something like this:
7071 \c ; and then, further down...
7073 \c push dword [myint] ; one of my integer variables
7074 \c push dword mystring ; pointer into my data segment
7076 \c add esp,byte 8 ; `byte' saves space
7078 \c ; then those data items...
7083 \c mystring db 'This number -> %d <- should be 1234',10,0
7085 This piece of code is the assembly equivalent of the C code
7087 \c int myint = 1234;
7088 \c printf("This number -> %d <- should be 1234\n", myint);
7091 \S{32cdata} Accessing Data Items
7093 To get at the contents of C variables, or to declare variables which
7094 C can access, you need only declare the names as \c{GLOBAL} or
7095 \c{EXTERN}. (Again, the names require leading underscores, as stated
7096 in \k{32cunder}.) Thus, a C variable declared as \c{int i} can be
7097 accessed from assembler as
7102 And to declare your own integer variable which C programs can access
7103 as \c{extern int j}, you do this (making sure you are assembling in
7104 the \c{_DATA} segment, if necessary):
7109 To access a C array, you need to know the size of the components of
7110 the array. For example, \c{int} variables are four bytes long, so if
7111 a C program declares an array as \c{int a[10]}, you can access
7112 \c{a[3]} by coding \c{mov ax,[_a+12]}. (The byte offset 12 is obtained
7113 by multiplying the desired array index, 3, by the size of the array
7114 element, 4.) The sizes of the C base types in 32-bit compilers are:
7115 1 for \c{char}, 2 for \c{short}, 4 for \c{int}, \c{long} and
7116 \c{float}, and 8 for \c{double}. Pointers, being 32-bit addresses,
7117 are also 4 bytes long.
7119 To access a C \i{data structure}, you need to know the offset from
7120 the base of the structure to the field you are interested in. You
7121 can either do this by converting the C structure definition into a
7122 NASM structure definition (using \c{STRUC}), or by calculating the
7123 one offset and using just that.
7125 To do either of these, you should read your C compiler's manual to
7126 find out how it organizes data structures. NASM gives no special
7127 alignment to structure members in its own \i\c{STRUC} macro, so you
7128 have to specify alignment yourself if the C compiler generates it.
7129 Typically, you might find that a structure like
7136 might be eight bytes long rather than five, since the \c{int} field
7137 would be aligned to a four-byte boundary. However, this sort of
7138 feature is sometimes a configurable option in the C compiler, either
7139 using command-line options or \c{#pragma} lines, so you have to find
7140 out how your own compiler does it.
7143 \S{32cmacro} \i\c{c32.mac}: Helper Macros for the 32-bit C Interface
7145 Included in the NASM archives, in the \I{misc directory}\c{misc}
7146 directory, is a file \c{c32.mac} of macros. It defines three macros:
7147 \i\c{proc}, \i\c{arg} and \i\c{endproc}. These are intended to be
7148 used for C-style procedure definitions, and they automate a lot of
7149 the work involved in keeping track of the calling convention.
7151 An example of an assembly function using the macro set is given
7158 \c mov eax,[ebp + %$i]
7159 \c mov ebx,[ebp + %$j]
7164 This defines \c{_proc32} to be a procedure taking two arguments, the
7165 first (\c{i}) an integer and the second (\c{j}) a pointer to an
7166 integer. It returns \c{i + *j}.
7168 Note that the \c{arg} macro has an \c{EQU} as the first line of its
7169 expansion, and since the label before the macro call gets prepended
7170 to the first line of the expanded macro, the \c{EQU} works, defining
7171 \c{%$i} to be an offset from \c{BP}. A context-local variable is
7172 used, local to the context pushed by the \c{proc} macro and popped
7173 by the \c{endproc} macro, so that the same argument name can be used
7174 in later procedures. Of course, you don't \e{have} to do that.
7176 \c{arg} can take an optional parameter, giving the size of the
7177 argument. If no size is given, 4 is assumed, since it is likely that
7178 many function parameters will be of type \c{int} or pointers.
7181 \H{picdll} Writing NetBSD/FreeBSD/OpenBSD and Linux/ELF \i{Shared
7184 \c{ELF} replaced the older \c{a.out} object file format under Linux
7185 because it contains support for \i{position-independent code}
7186 (\i{PIC}), which makes writing shared libraries much easier. NASM
7187 supports the \c{ELF} position-independent code features, so you can
7188 write Linux \c{ELF} shared libraries in NASM.
7190 \i{NetBSD}, and its close cousins \i{FreeBSD} and \i{OpenBSD}, take
7191 a different approach by hacking PIC support into the \c{a.out}
7192 format. NASM supports this as the \i\c{aoutb} output format, so you
7193 can write \i{BSD} shared libraries in NASM too.
7195 The operating system loads a PIC shared library by memory-mapping
7196 the library file at an arbitrarily chosen point in the address space
7197 of the running process. The contents of the library's code section
7198 must therefore not depend on where it is loaded in memory.
7200 Therefore, you cannot get at your variables by writing code like
7203 \c mov eax,[myvar] ; WRONG
7205 Instead, the linker provides an area of memory called the
7206 \i\e{global offset table}, or \i{GOT}; the GOT is situated at a
7207 constant distance from your library's code, so if you can find out
7208 where your library is loaded (which is typically done using a
7209 \c{CALL} and \c{POP} combination), you can obtain the address of the
7210 GOT, and you can then load the addresses of your variables out of
7211 linker-generated entries in the GOT.
7213 The \e{data} section of a PIC shared library does not have these
7214 restrictions: since the data section is writable, it has to be
7215 copied into memory anyway rather than just paged in from the library
7216 file, so as long as it's being copied it can be relocated too. So
7217 you can put ordinary types of relocation in the data section without
7218 too much worry (but see \k{picglobal} for a caveat).
7221 \S{picgot} Obtaining the Address of the GOT
7223 Each code module in your shared library should define the GOT as an
7226 \c extern _GLOBAL_OFFSET_TABLE_ ; in ELF
7227 \c extern __GLOBAL_OFFSET_TABLE_ ; in BSD a.out
7229 At the beginning of any function in your shared library which plans
7230 to access your data or BSS sections, you must first calculate the
7231 address of the GOT. This is typically done by writing the function
7240 \c add ebx,_GLOBAL_OFFSET_TABLE_+$$-.get_GOT wrt ..gotpc
7242 \c ; the function body comes here
7249 (For BSD, again, the symbol \c{_GLOBAL_OFFSET_TABLE} requires a
7250 second leading underscore.)
7252 The first two lines of this function are simply the standard C
7253 prologue to set up a stack frame, and the last three lines are
7254 standard C function epilogue. The third line, and the fourth to last
7255 line, save and restore the \c{EBX} register, because PIC shared
7256 libraries use this register to store the address of the GOT.
7258 The interesting bit is the \c{CALL} instruction and the following
7259 two lines. The \c{CALL} and \c{POP} combination obtains the address
7260 of the label \c{.get_GOT}, without having to know in advance where
7261 the program was loaded (since the \c{CALL} instruction is encoded
7262 relative to the current position). The \c{ADD} instruction makes use
7263 of one of the special PIC relocation types: \i{GOTPC relocation}.
7264 With the \i\c{WRT ..gotpc} qualifier specified, the symbol
7265 referenced (here \c{_GLOBAL_OFFSET_TABLE_}, the special symbol
7266 assigned to the GOT) is given as an offset from the beginning of the
7267 section. (Actually, \c{ELF} encodes it as the offset from the operand
7268 field of the \c{ADD} instruction, but NASM simplifies this
7269 deliberately, so you do things the same way for both \c{ELF} and
7270 \c{BSD}.) So the instruction then \e{adds} the beginning of the section,
7271 to get the real address of the GOT, and subtracts the value of
7272 \c{.get_GOT} which it knows is in \c{EBX}. Therefore, by the time
7273 that instruction has finished, \c{EBX} contains the address of the GOT.
7275 If you didn't follow that, don't worry: it's never necessary to
7276 obtain the address of the GOT by any other means, so you can put
7277 those three instructions into a macro and safely ignore them:
7284 \c add ebx,_GLOBAL_OFFSET_TABLE_+$$-%%getgot wrt ..gotpc
7288 \S{piclocal} Finding Your Local Data Items
7290 Having got the GOT, you can then use it to obtain the addresses of
7291 your data items. Most variables will reside in the sections you have
7292 declared; they can be accessed using the \I{GOTOFF
7293 relocation}\c{..gotoff} special \I\c{WRT ..gotoff}\c{WRT} type. The
7294 way this works is like this:
7296 \c lea eax,[ebx+myvar wrt ..gotoff]
7298 The expression \c{myvar wrt ..gotoff} is calculated, when the shared
7299 library is linked, to be the offset to the local variable \c{myvar}
7300 from the beginning of the GOT. Therefore, adding it to \c{EBX} as
7301 above will place the real address of \c{myvar} in \c{EAX}.
7303 If you declare variables as \c{GLOBAL} without specifying a size for
7304 them, they are shared between code modules in the library, but do
7305 not get exported from the library to the program that loaded it.
7306 They will still be in your ordinary data and BSS sections, so you
7307 can access them in the same way as local variables, using the above
7308 \c{..gotoff} mechanism.
7310 Note that due to a peculiarity of the way BSD \c{a.out} format
7311 handles this relocation type, there must be at least one non-local
7312 symbol in the same section as the address you're trying to access.
7315 \S{picextern} Finding External and Common Data Items
7317 If your library needs to get at an external variable (external to
7318 the \e{library}, not just to one of the modules within it), you must
7319 use the \I{GOT relocations}\I\c{WRT ..got}\c{..got} type to get at
7320 it. The \c{..got} type, instead of giving you the offset from the
7321 GOT base to the variable, gives you the offset from the GOT base to
7322 a GOT \e{entry} containing the address of the variable. The linker
7323 will set up this GOT entry when it builds the library, and the
7324 dynamic linker will place the correct address in it at load time. So
7325 to obtain the address of an external variable \c{extvar} in \c{EAX},
7328 \c mov eax,[ebx+extvar wrt ..got]
7330 This loads the address of \c{extvar} out of an entry in the GOT. The
7331 linker, when it builds the shared library, collects together every
7332 relocation of type \c{..got}, and builds the GOT so as to ensure it
7333 has every necessary entry present.
7335 Common variables must also be accessed in this way.
7338 \S{picglobal} Exporting Symbols to the Library User
7340 If you want to export symbols to the user of the library, you have
7341 to declare whether they are functions or data, and if they are data,
7342 you have to give the size of the data item. This is because the
7343 dynamic linker has to build \I{PLT}\i{procedure linkage table}
7344 entries for any exported functions, and also moves exported data
7345 items away from the library's data section in which they were
7348 So to export a function to users of the library, you must use
7350 \c global func:function ; declare it as a function
7356 And to export a data item such as an array, you would have to code
7358 \c global array:data array.end-array ; give the size too
7363 Be careful: If you export a variable to the library user, by
7364 declaring it as \c{GLOBAL} and supplying a size, the variable will
7365 end up living in the data section of the main program, rather than
7366 in your library's data section, where you declared it. So you will
7367 have to access your own global variable with the \c{..got} mechanism
7368 rather than \c{..gotoff}, as if it were external (which,
7369 effectively, it has become).
7371 Equally, if you need to store the address of an exported global in
7372 one of your data sections, you can't do it by means of the standard
7375 \c dataptr: dd global_data_item ; WRONG
7377 NASM will interpret this code as an ordinary relocation, in which
7378 \c{global_data_item} is merely an offset from the beginning of the
7379 \c{.data} section (or whatever); so this reference will end up
7380 pointing at your data section instead of at the exported global
7381 which resides elsewhere.
7383 Instead of the above code, then, you must write
7385 \c dataptr: dd global_data_item wrt ..sym
7387 which makes use of the special \c{WRT} type \I\c{WRT ..sym}\c{..sym}
7388 to instruct NASM to search the symbol table for a particular symbol
7389 at that address, rather than just relocating by section base.
7391 Either method will work for functions: referring to one of your
7392 functions by means of
7394 \c funcptr: dd my_function
7396 will give the user the address of the code you wrote, whereas
7398 \c funcptr: dd my_function wrt ..sym
7400 will give the address of the procedure linkage table for the
7401 function, which is where the calling program will \e{believe} the
7402 function lives. Either address is a valid way to call the function.
7405 \S{picproc} Calling Procedures Outside the Library
7407 Calling procedures outside your shared library has to be done by
7408 means of a \i\e{procedure linkage table}, or \i{PLT}. The PLT is
7409 placed at a known offset from where the library is loaded, so the
7410 library code can make calls to the PLT in a position-independent
7411 way. Within the PLT there is code to jump to offsets contained in
7412 the GOT, so function calls to other shared libraries or to routines
7413 in the main program can be transparently passed off to their real
7416 To call an external routine, you must use another special PIC
7417 relocation type, \I{PLT relocations}\i\c{WRT ..plt}. This is much
7418 easier than the GOT-based ones: you simply replace calls such as
7419 \c{CALL printf} with the PLT-relative version \c{CALL printf WRT
7423 \S{link} Generating the Library File
7425 Having written some code modules and assembled them to \c{.o} files,
7426 you then generate your shared library with a command such as
7428 \c ld -shared -o library.so module1.o module2.o # for ELF
7429 \c ld -Bshareable -o library.so module1.o module2.o # for BSD
7431 For ELF, if your shared library is going to reside in system
7432 directories such as \c{/usr/lib} or \c{/lib}, it is usually worth
7433 using the \i\c{-soname} flag to the linker, to store the final
7434 library file name, with a version number, into the library:
7436 \c ld -shared -soname library.so.1 -o library.so.1.2 *.o
7438 You would then copy \c{library.so.1.2} into the library directory,
7439 and create \c{library.so.1} as a symbolic link to it.
7442 \C{mixsize} Mixing 16 and 32 Bit Code
7444 This chapter tries to cover some of the issues, largely related to
7445 unusual forms of addressing and jump instructions, encountered when
7446 writing operating system code such as protected-mode initialisation
7447 routines, which require code that operates in mixed segment sizes,
7448 such as code in a 16-bit segment trying to modify data in a 32-bit
7449 one, or jumps between different-size segments.
7452 \H{mixjump} Mixed-Size Jumps\I{jumps, mixed-size}
7454 \I{operating system, writing}\I{writing operating systems}The most
7455 common form of \i{mixed-size instruction} is the one used when
7456 writing a 32-bit OS: having done your setup in 16-bit mode, such as
7457 loading the kernel, you then have to boot it by switching into
7458 protected mode and jumping to the 32-bit kernel start address. In a
7459 fully 32-bit OS, this tends to be the \e{only} mixed-size
7460 instruction you need, since everything before it can be done in pure
7461 16-bit code, and everything after it can be pure 32-bit.
7463 This jump must specify a 48-bit far address, since the target
7464 segment is a 32-bit one. However, it must be assembled in a 16-bit
7465 segment, so just coding, for example,
7467 \c jmp 0x1234:0x56789ABC ; wrong!
7469 will not work, since the offset part of the address will be
7470 truncated to \c{0x9ABC} and the jump will be an ordinary 16-bit far
7473 The Linux kernel setup code gets round the inability of \c{as86} to
7474 generate the required instruction by coding it manually, using
7475 \c{DB} instructions. NASM can go one better than that, by actually
7476 generating the right instruction itself. Here's how to do it right:
7478 \c jmp dword 0x1234:0x56789ABC ; right
7480 \I\c{JMP DWORD}The \c{DWORD} prefix (strictly speaking, it should
7481 come \e{after} the colon, since it is declaring the \e{offset} field
7482 to be a doubleword; but NASM will accept either form, since both are
7483 unambiguous) forces the offset part to be treated as far, in the
7484 assumption that you are deliberately writing a jump from a 16-bit
7485 segment to a 32-bit one.
7487 You can do the reverse operation, jumping from a 32-bit segment to a
7488 16-bit one, by means of the \c{WORD} prefix:
7490 \c jmp word 0x8765:0x4321 ; 32 to 16 bit
7492 If the \c{WORD} prefix is specified in 16-bit mode, or the \c{DWORD}
7493 prefix in 32-bit mode, they will be ignored, since each is
7494 explicitly forcing NASM into a mode it was in anyway.
7497 \H{mixaddr} Addressing Between Different-Size Segments\I{addressing,
7498 mixed-size}\I{mixed-size addressing}
7500 If your OS is mixed 16 and 32-bit, or if you are writing a DOS
7501 extender, you are likely to have to deal with some 16-bit segments
7502 and some 32-bit ones. At some point, you will probably end up
7503 writing code in a 16-bit segment which has to access data in a
7504 32-bit segment, or vice versa.
7506 If the data you are trying to access in a 32-bit segment lies within
7507 the first 64K of the segment, you may be able to get away with using
7508 an ordinary 16-bit addressing operation for the purpose; but sooner
7509 or later, you will want to do 32-bit addressing from 16-bit mode.
7511 The easiest way to do this is to make sure you use a register for
7512 the address, since any effective address containing a 32-bit
7513 register is forced to be a 32-bit address. So you can do
7515 \c mov eax,offset_into_32_bit_segment_specified_by_fs
7516 \c mov dword [fs:eax],0x11223344
7518 This is fine, but slightly cumbersome (since it wastes an
7519 instruction and a register) if you already know the precise offset
7520 you are aiming at. The x86 architecture does allow 32-bit effective
7521 addresses to specify nothing but a 4-byte offset, so why shouldn't
7522 NASM be able to generate the best instruction for the purpose?
7524 It can. As in \k{mixjump}, you need only prefix the address with the
7525 \c{DWORD} keyword, and it will be forced to be a 32-bit address:
7527 \c mov dword [fs:dword my_offset],0x11223344
7529 Also as in \k{mixjump}, NASM is not fussy about whether the
7530 \c{DWORD} prefix comes before or after the segment override, so
7531 arguably a nicer-looking way to code the above instruction is
7533 \c mov dword [dword fs:my_offset],0x11223344
7535 Don't confuse the \c{DWORD} prefix \e{outside} the square brackets,
7536 which controls the size of the data stored at the address, with the
7537 one \c{inside} the square brackets which controls the length of the
7538 address itself. The two can quite easily be different:
7540 \c mov word [dword 0x12345678],0x9ABC
7542 This moves 16 bits of data to an address specified by a 32-bit
7545 You can also specify \c{WORD} or \c{DWORD} prefixes along with the
7546 \c{FAR} prefix to indirect far jumps or calls. For example:
7548 \c call dword far [fs:word 0x4321]
7550 This instruction contains an address specified by a 16-bit offset;
7551 it loads a 48-bit far pointer from that (16-bit segment and 32-bit
7552 offset), and calls that address.
7555 \H{mixother} Other Mixed-Size Instructions
7557 The other way you might want to access data might be using the
7558 string instructions (\c{LODSx}, \c{STOSx} and so on) or the
7559 \c{XLATB} instruction. These instructions, since they take no
7560 parameters, might seem to have no easy way to make them perform
7561 32-bit addressing when assembled in a 16-bit segment.
7563 This is the purpose of NASM's \i\c{a16}, \i\c{a32} and \i\c{a64} prefixes. If
7564 you are coding \c{LODSB} in a 16-bit segment but it is supposed to
7565 be accessing a string in a 32-bit segment, you should load the
7566 desired address into \c{ESI} and then code
7570 The prefix forces the addressing size to 32 bits, meaning that
7571 \c{LODSB} loads from \c{[DS:ESI]} instead of \c{[DS:SI]}. To access
7572 a string in a 16-bit segment when coding in a 32-bit one, the
7573 corresponding \c{a16} prefix can be used.
7575 The \c{a16}, \c{a32} and \c{a64} prefixes can be applied to any instruction
7576 in NASM's instruction table, but most of them can generate all the
7577 useful forms without them. The prefixes are necessary only for
7578 instructions with implicit addressing:
7579 \# \c{CMPSx} (\k{insCMPSB}),
7580 \# \c{SCASx} (\k{insSCASB}), \c{LODSx} (\k{insLODSB}), \c{STOSx}
7581 \# (\k{insSTOSB}), \c{MOVSx} (\k{insMOVSB}), \c{INSx} (\k{insINSB}),
7582 \# \c{OUTSx} (\k{insOUTSB}), and \c{XLATB} (\k{insXLATB}).
7583 \c{CMPSx}, \c{SCASx}, \c{LODSx}, \c{STOSx}, \c{MOVSx}, \c{INSx},
7584 \c{OUTSx}, and \c{XLATB}.
7586 various push and pop instructions (\c{PUSHA} and \c{POPF} as well as
7587 the more usual \c{PUSH} and \c{POP}) can accept \c{a16}, \c{a32} or \c{a64}
7588 prefixes to force a particular one of \c{SP}, \c{ESP} or \c{RSP} to be used
7589 as a stack pointer, in case the stack segment in use is a different
7590 size from the code segment.
7592 \c{PUSH} and \c{POP}, when applied to segment registers in 32-bit
7593 mode, also have the slightly odd behaviour that they push and pop 4
7594 bytes at a time, of which the top two are ignored and the bottom two
7595 give the value of the segment register being manipulated. To force
7596 the 16-bit behaviour of segment-register push and pop instructions,
7597 you can use the operand-size prefix \i\c{o16}:
7602 This code saves a doubleword of stack space by fitting two segment
7603 registers into the space which would normally be consumed by pushing
7606 (You can also use the \i\c{o32} prefix to force the 32-bit behaviour
7607 when in 16-bit mode, but this seems less useful.)
7610 \C{64bit} Writing 64-bit Code (Unix, Win64)
7612 This chapter attempts to cover some of the common issues involved when
7613 writing 64-bit code, to run under \i{Win64} or Unix. It covers how to
7614 write assembly code to interface with 64-bit C routines, and how to
7615 write position-independent code for shared libraries.
7617 All 64-bit code uses a flat memory model, since segmentation is not
7618 available in 64-bit mode. The one exception is the \c{FS} and \c{GS}
7619 registers, which still add their bases.
7621 Position independence in 64-bit mode is significantly simpler, since
7622 the processor supports \c{RIP}-relative addressing directly; see the
7623 \c{REL} keyword (\k{effaddr}). On most 64-bit platforms, it is
7624 probably desirable to make that the default, using the directive
7625 \c{DEFAULT REL} (\k{default}).
7627 64-bit programming is relatively similar to 32-bit programming, but
7628 of course pointers are 64 bits long; additionally, all existing
7629 platforms pass arguments in registers rather than on the stack.
7630 Furthermore, 64-bit platforms use SSE2 by default for floating point.
7631 Please see the ABI documentation for your platform.
7633 64-bit platforms differ in the sizes of the fundamental datatypes, not
7634 just from 32-bit platforms but from each other. If a specific size
7635 data type is desired, it is probably best to use the types defined in
7636 the Standard C header \c{<inttypes.h>}.
7638 In 64-bit mode, the default instruction size is still 32 bits. When
7639 loading a value into a 32-bit register (but not an 8- or 16-bit
7640 register), the upper 32 bits of the corresponding 64-bit register are
7643 \H{reg64} Register Names in 64-bit Mode
7645 NASM uses the following names for general-purpose registers in 64-bit
7646 mode, for 8-, 16-, 32- and 64-bit references, respecitively:
7648 \c AL/AH, CL/CH, DL/DH, BL/BH, SPL, BPL, SIL, DIL, R8B-R15B
7649 \c AX, CX, DX, BX, SP, BP, SI, DI, R8W-R15W
7650 \c EAX, ECX, EDX, EBX, ESP, EBP, ESI, EDI, R8D-R15D
7651 \c RAX, RCX, RDX, RBX, RSP, RBP, RSI, RDI, R8-R15
7653 This is consistent with the AMD documentation and most other
7654 assemblers. The Intel documentation, however, uses the names
7655 \c{R8L-R15L} for 8-bit references to the higher registers. It is
7656 possible to use those names by definiting them as macros; similarly,
7657 if one wants to use numeric names for the low 8 registers, define them
7658 as macros. The standard macro package \c{altreg} (see \k{pkg_altreg})
7659 can be used for this purpose.
7661 \H{id64} Immediates and Displacements in 64-bit Mode
7663 In 64-bit mode, immediates and displacements are generally only 32
7664 bits wide. NASM will therefore truncate most displacements and
7665 immediates to 32 bits.
7667 The only instruction which takes a full \i{64-bit immediate} is:
7671 NASM will produce this instruction whenever the programmer uses
7672 \c{MOV} with an immediate into a 64-bit register. If this is not
7673 desirable, simply specify the equivalent 32-bit register, which will
7674 be automatically zero-extended by the processor, or specify the
7675 immediate as \c{DWORD}:
7677 \c mov rax,foo ; 64-bit immediate
7678 \c mov rax,qword foo ; (identical)
7679 \c mov eax,foo ; 32-bit immediate, zero-extended
7680 \c mov rax,dword foo ; 32-bit immediate, sign-extended
7682 The length of these instructions are 10, 5 and 7 bytes, respectively.
7684 The only instructions which take a full \I{64-bit displacement}64-bit
7685 \e{displacement} is loading or storing, using \c{MOV}, \c{AL}, \c{AX},
7686 \c{EAX} or \c{RAX} (but no other registers) to an absolute 64-bit address.
7687 Since this is a relatively rarely used instruction (64-bit code generally uses
7688 relative addressing), the programmer has to explicitly declare the
7689 displacement size as \c{QWORD}:
7693 \c mov eax,[foo] ; 32-bit absolute disp, sign-extended
7694 \c mov eax,[a32 foo] ; 32-bit absolute disp, zero-extended
7695 \c mov eax,[qword foo] ; 64-bit absolute disp
7699 \c mov eax,[foo] ; 32-bit relative disp
7700 \c mov eax,[a32 foo] ; d:o, address truncated to 32 bits(!)
7701 \c mov eax,[qword foo] ; error
7702 \c mov eax,[abs qword foo] ; 64-bit absolute disp
7704 A sign-extended absolute displacement can access from -2 GB to +2 GB;
7705 a zero-extended absolute displacement can access from 0 to 4 GB.
7707 \H{unix64} Interfacing to 64-bit C Programs (Unix)
7709 On Unix, the 64-bit ABI is defined by the document:
7711 \W{http://www.nasm.us/links/unix64abi}\c{http://www.nasm.us/links/unix64abi}
7713 Although written for AT&T-syntax assembly, the concepts apply equally
7714 well for NASM-style assembly. What follows is a simplified summary.
7716 The first six integer arguments (from the left) are passed in \c{RDI},
7717 \c{RSI}, \c{RDX}, \c{RCX}, \c{R8}, and \c{R9}, in that order.
7718 Additional integer arguments are passed on the stack. These
7719 registers, plus \c{RAX}, \c{R10} and \c{R11} are destroyed by function
7720 calls, and thus are available for use by the function without saving.
7722 Integer return values are passed in \c{RAX} and \c{RDX}, in that order.
7724 Floating point is done using SSE registers, except for \c{long
7725 double}. Floating-point arguments are passed in \c{XMM0} to \c{XMM7};
7726 return is \c{XMM0} and \c{XMM1}. \c{long double} are passed on the
7727 stack, and returned in \c{ST0} and \c{ST1}.
7729 All SSE and x87 registers are destroyed by function calls.
7731 On 64-bit Unix, \c{long} is 64 bits.
7733 Integer and SSE register arguments are counted separately, so for the case of
7735 \c void foo(long a, double b, int c)
7737 \c{a} is passed in \c{RDI}, \c{b} in \c{XMM0}, and \c{c} in \c{ESI}.
7739 \H{win64} Interfacing to 64-bit C Programs (Win64)
7741 The Win64 ABI is described at:
7743 \W{http://www.nasm.us/links/win64abi}\c{http://www.nasm.us/links/win64abi}
7745 What follows is a simplified summary.
7747 The first four integer arguments are passed in \c{RCX}, \c{RDX},
7748 \c{R8} and \c{R9}, in that order. Additional integer arguments are
7749 passed on the stack. These registers, plus \c{RAX}, \c{R10} and
7750 \c{R11} are destroyed by function calls, and thus are available for
7751 use by the function without saving.
7753 Integer return values are passed in \c{RAX} only.
7755 Floating point is done using SSE registers, except for \c{long
7756 double}. Floating-point arguments are passed in \c{XMM0} to \c{XMM3};
7757 return is \c{XMM0} only.
7759 On Win64, \c{long} is 32 bits; \c{long long} or \c{_int64} is 64 bits.
7761 Integer and SSE register arguments are counted together, so for the case of
7763 \c void foo(long long a, double b, int c)
7765 \c{a} is passed in \c{RCX}, \c{b} in \c{XMM1}, and \c{c} in \c{R8D}.
7767 \C{trouble} Troubleshooting
7769 This chapter describes some of the common problems that users have
7770 been known to encounter with NASM, and answers them. It also gives
7771 instructions for reporting bugs in NASM if you find a difficulty
7772 that isn't listed here.
7775 \H{problems} Common Problems
7777 \S{inefficient} NASM Generates \i{Inefficient Code}
7779 We sometimes get `bug' reports about NASM generating inefficient, or
7780 even `wrong', code on instructions such as \c{ADD ESP,8}. This is a
7781 deliberate design feature, connected to predictability of output:
7782 NASM, on seeing \c{ADD ESP,8}, will generate the form of the
7783 instruction which leaves room for a 32-bit offset. You need to code
7784 \I\c{BYTE}\c{ADD ESP,BYTE 8} if you want the space-efficient form of
7785 the instruction. This isn't a bug, it's user error: if you prefer to
7786 have NASM produce the more efficient code automatically enable
7787 optimization with the \c{-O} option (see \k{opt-O}).
7790 \S{jmprange} My Jumps are Out of Range\I{out of range, jumps}
7792 Similarly, people complain that when they issue \i{conditional
7793 jumps} (which are \c{SHORT} by default) that try to jump too far,
7794 NASM reports `short jump out of range' instead of making the jumps
7797 This, again, is partly a predictability issue, but in fact has a
7798 more practical reason as well. NASM has no means of being told what
7799 type of processor the code it is generating will be run on; so it
7800 cannot decide for itself that it should generate \i\c{Jcc NEAR} type
7801 instructions, because it doesn't know that it's working for a 386 or
7802 above. Alternatively, it could replace the out-of-range short
7803 \c{JNE} instruction with a very short \c{JE} instruction that jumps
7804 over a \c{JMP NEAR}; this is a sensible solution for processors
7805 below a 386, but hardly efficient on processors which have good
7806 branch prediction \e{and} could have used \c{JNE NEAR} instead. So,
7807 once again, it's up to the user, not the assembler, to decide what
7808 instructions should be generated. See \k{opt-O}.
7811 \S{proborg} \i\c{ORG} Doesn't Work
7813 People writing \i{boot sector} programs in the \c{bin} format often
7814 complain that \c{ORG} doesn't work the way they'd like: in order to
7815 place the \c{0xAA55} signature word at the end of a 512-byte boot
7816 sector, people who are used to MASM tend to code
7820 \c ; some boot sector code
7825 This is not the intended use of the \c{ORG} directive in NASM, and
7826 will not work. The correct way to solve this problem in NASM is to
7827 use the \i\c{TIMES} directive, like this:
7831 \c ; some boot sector code
7833 \c TIMES 510-($-$$) DB 0
7836 The \c{TIMES} directive will insert exactly enough zero bytes into
7837 the output to move the assembly point up to 510. This method also
7838 has the advantage that if you accidentally fill your boot sector too
7839 full, NASM will catch the problem at assembly time and report it, so
7840 you won't end up with a boot sector that you have to disassemble to
7841 find out what's wrong with it.
7844 \S{probtimes} \i\c{TIMES} Doesn't Work
7846 The other common problem with the above code is people who write the
7851 by reasoning that \c{$} should be a pure number, just like 510, so
7852 the difference between them is also a pure number and can happily be
7855 NASM is a \e{modular} assembler: the various component parts are
7856 designed to be easily separable for re-use, so they don't exchange
7857 information unnecessarily. In consequence, the \c{bin} output
7858 format, even though it has been told by the \c{ORG} directive that
7859 the \c{.text} section should start at 0, does not pass that
7860 information back to the expression evaluator. So from the
7861 evaluator's point of view, \c{$} isn't a pure number: it's an offset
7862 from a section base. Therefore the difference between \c{$} and 510
7863 is also not a pure number, but involves a section base. Values
7864 involving section bases cannot be passed as arguments to \c{TIMES}.
7866 The solution, as in the previous section, is to code the \c{TIMES}
7869 \c TIMES 510-($-$$) DB 0
7871 in which \c{$} and \c{$$} are offsets from the same section base,
7872 and so their difference is a pure number. This will solve the
7873 problem and generate sensible code.
7876 \H{bugs} \i{Bugs}\I{reporting bugs}
7878 We have never yet released a version of NASM with any \e{known}
7879 bugs. That doesn't usually stop there being plenty we didn't know
7880 about, though. Any that you find should be reported firstly via the
7882 \W{http://www.nasm.us/}\c{http://www.nasm.us/}
7883 (click on "Bug Tracker"), or if that fails then through one of the
7884 contacts in \k{contact}.
7886 Please read \k{qstart} first, and don't report the bug if it's
7887 listed in there as a deliberate feature. (If you think the feature
7888 is badly thought out, feel free to send us reasons why you think it
7889 should be changed, but don't just send us mail saying `This is a
7890 bug' if the documentation says we did it on purpose.) Then read
7891 \k{problems}, and don't bother reporting the bug if it's listed
7894 If you do report a bug, \e{please} give us all of the following
7897 \b What operating system you're running NASM under. DOS, Linux,
7898 NetBSD, Win16, Win32, VMS (I'd be impressed), whatever.
7900 \b If you're running NASM under DOS or Win32, tell us whether you've
7901 compiled your own executable from the DOS source archive, or whether
7902 you were using the standard distribution binaries out of the
7903 archive. If you were using a locally built executable, try to
7904 reproduce the problem using one of the standard binaries, as this
7905 will make it easier for us to reproduce your problem prior to fixing
7908 \b Which version of NASM you're using, and exactly how you invoked
7909 it. Give us the precise command line, and the contents of the
7910 \c{NASMENV} environment variable if any.
7912 \b Which versions of any supplementary programs you're using, and
7913 how you invoked them. If the problem only becomes visible at link
7914 time, tell us what linker you're using, what version of it you've
7915 got, and the exact linker command line. If the problem involves
7916 linking against object files generated by a compiler, tell us what
7917 compiler, what version, and what command line or options you used.
7918 (If you're compiling in an IDE, please try to reproduce the problem
7919 with the command-line version of the compiler.)
7921 \b If at all possible, send us a NASM source file which exhibits the
7922 problem. If this causes copyright problems (e.g. you can only
7923 reproduce the bug in restricted-distribution code) then bear in mind
7924 the following two points: firstly, we guarantee that any source code
7925 sent to us for the purposes of debugging NASM will be used \e{only}
7926 for the purposes of debugging NASM, and that we will delete all our
7927 copies of it as soon as we have found and fixed the bug or bugs in
7928 question; and secondly, we would prefer \e{not} to be mailed large
7929 chunks of code anyway. The smaller the file, the better. A
7930 three-line sample file that does nothing useful \e{except}
7931 demonstrate the problem is much easier to work with than a
7932 fully fledged ten-thousand-line program. (Of course, some errors
7933 \e{do} only crop up in large files, so this may not be possible.)
7935 \b A description of what the problem actually \e{is}. `It doesn't
7936 work' is \e{not} a helpful description! Please describe exactly what
7937 is happening that shouldn't be, or what isn't happening that should.
7938 Examples might be: `NASM generates an error message saying Line 3
7939 for an error that's actually on Line 5'; `NASM generates an error
7940 message that I believe it shouldn't be generating at all'; `NASM
7941 fails to generate an error message that I believe it \e{should} be
7942 generating'; `the object file produced from this source code crashes
7943 my linker'; `the ninth byte of the output file is 66 and I think it
7944 should be 77 instead'.
7946 \b If you believe the output file from NASM to be faulty, send it to
7947 us. That allows us to determine whether our own copy of NASM
7948 generates the same file, or whether the problem is related to
7949 portability issues between our development platforms and yours. We
7950 can handle binary files mailed to us as MIME attachments, uuencoded,
7951 and even BinHex. Alternatively, we may be able to provide an FTP
7952 site you can upload the suspect files to; but mailing them is easier
7955 \b Any other information or data files that might be helpful. If,
7956 for example, the problem involves NASM failing to generate an object
7957 file while TASM can generate an equivalent file without trouble,
7958 then send us \e{both} object files, so we can see what TASM is doing
7959 differently from us.
7962 \A{ndisasm} \i{Ndisasm}
7964 The Netwide Disassembler, NDISASM
7966 \H{ndisintro} Introduction
7969 The Netwide Disassembler is a small companion program to the Netwide
7970 Assembler, NASM. It seemed a shame to have an x86 assembler,
7971 complete with a full instruction table, and not make as much use of
7972 it as possible, so here's a disassembler which shares the
7973 instruction table (and some other bits of code) with NASM.
7975 The Netwide Disassembler does nothing except to produce
7976 disassemblies of \e{binary} source files. NDISASM does not have any
7977 understanding of object file formats, like \c{objdump}, and it will
7978 not understand \c{DOS .EXE} files like \c{debug} will. It just
7982 \H{ndisstart} Getting Started: Installation
7984 See \k{install} for installation instructions. NDISASM, like NASM,
7985 has a \c{man page} which you may want to put somewhere useful, if you
7986 are on a Unix system.
7989 \H{ndisrun} Running NDISASM
7991 To disassemble a file, you will typically use a command of the form
7993 \c ndisasm -b {16|32|64} filename
7995 NDISASM can disassemble 16-, 32- or 64-bit code equally easily,
7996 provided of course that you remember to specify which it is to work
7997 with. If no \i\c{-b} switch is present, NDISASM works in 16-bit mode
7998 by default. The \i\c{-u} switch (for USE32) also invokes 32-bit mode.
8000 Two more command line options are \i\c{-r} which reports the version
8001 number of NDISASM you are running, and \i\c{-h} which gives a short
8002 summary of command line options.
8005 \S{ndiscom} COM Files: Specifying an Origin
8007 To disassemble a \c{DOS .COM} file correctly, a disassembler must assume
8008 that the first instruction in the file is loaded at address \c{0x100},
8009 rather than at zero. NDISASM, which assumes by default that any file
8010 you give it is loaded at zero, will therefore need to be informed of
8013 The \i\c{-o} option allows you to declare a different origin for the
8014 file you are disassembling. Its argument may be expressed in any of
8015 the NASM numeric formats: decimal by default, if it begins with `\c{$}'
8016 or `\c{0x}' or ends in `\c{H}' it's \c{hex}, if it ends in `\c{Q}' it's
8017 \c{octal}, and if it ends in `\c{B}' it's \c{binary}.
8019 Hence, to disassemble a \c{.COM} file:
8021 \c ndisasm -o100h filename.com
8026 \S{ndissync} Code Following Data: Synchronisation
8028 Suppose you are disassembling a file which contains some data which
8029 isn't machine code, and \e{then} contains some machine code. NDISASM
8030 will faithfully plough through the data section, producing machine
8031 instructions wherever it can (although most of them will look
8032 bizarre, and some may have unusual prefixes, e.g. `\c{FS OR AX,0x240A}'),
8033 and generating `DB' instructions ever so often if it's totally stumped.
8034 Then it will reach the code section.
8036 Supposing NDISASM has just finished generating a strange machine
8037 instruction from part of the data section, and its file position is
8038 now one byte \e{before} the beginning of the code section. It's
8039 entirely possible that another spurious instruction will get
8040 generated, starting with the final byte of the data section, and
8041 then the correct first instruction in the code section will not be
8042 seen because the starting point skipped over it. This isn't really
8045 To avoid this, you can specify a `\i\c{synchronisation}' point, or indeed
8046 as many synchronisation points as you like (although NDISASM can
8047 only handle 2147483647 sync points internally). The definition of a sync
8048 point is this: NDISASM guarantees to hit sync points exactly during
8049 disassembly. If it is thinking about generating an instruction which
8050 would cause it to jump over a sync point, it will discard that
8051 instruction and output a `\c{db}' instead. So it \e{will} start
8052 disassembly exactly from the sync point, and so you \e{will} see all
8053 the instructions in your code section.
8055 Sync points are specified using the \i\c{-s} option: they are measured
8056 in terms of the program origin, not the file position. So if you
8057 want to synchronize after 32 bytes of a \c{.COM} file, you would have to
8060 \c ndisasm -o100h -s120h file.com
8064 \c ndisasm -o100h -s20h file.com
8066 As stated above, you can specify multiple sync markers if you need
8067 to, just by repeating the \c{-s} option.
8070 \S{ndisisync} Mixed Code and Data: Automatic (Intelligent) Synchronisation
8073 Suppose you are disassembling the boot sector of a \c{DOS} floppy (maybe
8074 it has a virus, and you need to understand the virus so that you
8075 know what kinds of damage it might have done you). Typically, this
8076 will contain a \c{JMP} instruction, then some data, then the rest of the
8077 code. So there is a very good chance of NDISASM being \e{misaligned}
8078 when the data ends and the code begins. Hence a sync point is
8081 On the other hand, why should you have to specify the sync point
8082 manually? What you'd do in order to find where the sync point would
8083 be, surely, would be to read the \c{JMP} instruction, and then to use
8084 its target address as a sync point. So can NDISASM do that for you?
8086 The answer, of course, is yes: using either of the synonymous
8087 switches \i\c{-a} (for automatic sync) or \i\c{-i} (for intelligent
8088 sync) will enable \c{auto-sync} mode. Auto-sync mode automatically
8089 generates a sync point for any forward-referring PC-relative jump or
8090 call instruction that NDISASM encounters. (Since NDISASM is one-pass,
8091 if it encounters a PC-relative jump whose target has already been
8092 processed, there isn't much it can do about it...)
8094 Only PC-relative jumps are processed, since an absolute jump is
8095 either through a register (in which case NDISASM doesn't know what
8096 the register contains) or involves a segment address (in which case
8097 the target code isn't in the same segment that NDISASM is working
8098 in, and so the sync point can't be placed anywhere useful).
8100 For some kinds of file, this mechanism will automatically put sync
8101 points in all the right places, and save you from having to place
8102 any sync points manually. However, it should be stressed that
8103 auto-sync mode is \e{not} guaranteed to catch all the sync points, and
8104 you may still have to place some manually.
8106 Auto-sync mode doesn't prevent you from declaring manual sync
8107 points: it just adds automatically generated ones to the ones you
8108 provide. It's perfectly feasible to specify \c{-i} \e{and} some \c{-s}
8111 Another caveat with auto-sync mode is that if, by some unpleasant
8112 fluke, something in your data section should disassemble to a
8113 PC-relative call or jump instruction, NDISASM may obediently place a
8114 sync point in a totally random place, for example in the middle of
8115 one of the instructions in your code section. So you may end up with
8116 a wrong disassembly even if you use auto-sync. Again, there isn't
8117 much I can do about this. If you have problems, you'll have to use
8118 manual sync points, or use the \c{-k} option (documented below) to
8119 suppress disassembly of the data area.
8122 \S{ndisother} Other Options
8124 The \i\c{-e} option skips a header on the file, by ignoring the first N
8125 bytes. This means that the header is \e{not} counted towards the
8126 disassembly offset: if you give \c{-e10 -o10}, disassembly will start
8127 at byte 10 in the file, and this will be given offset 10, not 20.
8129 The \i\c{-k} option is provided with two comma-separated numeric
8130 arguments, the first of which is an assembly offset and the second
8131 is a number of bytes to skip. This \e{will} count the skipped bytes
8132 towards the assembly offset: its use is to suppress disassembly of a
8133 data section which wouldn't contain anything you wanted to see
8137 \H{ndisbugs} Bugs and Improvements
8139 There are no known bugs. However, any you find, with patches if
8140 possible, should be sent to
8141 \W{mailto:nasm-bugs@lists.sourceforge.net}\c{nasm-bugs@lists.sourceforge.net}, or to the
8143 \W{http://www.nasm.us/}\c{http://www.nasm.us/}
8144 and we'll try to fix them. Feel free to send contributions and
8145 new features as well.
8147 \A{inslist} \i{Instruction List}
8149 \H{inslistintro} Introduction
8151 The following sections show the instructions which NASM currently supports. For each
8152 instruction, there is a separate entry for each supported addressing mode. The third
8153 column shows the processor type in which the instruction was introduced and,
8154 when appropriate, one or more usage flags.
8158 \A{changelog} \i{NASM Version History}