1 \# --------------------------------------------------------------------------
3 \# Copyright 1996-2010 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{auxinfo}{This release is dedicated to the memory of Charles A. Crayne. We miss you, Chuck.}
43 \M{summary}{This file documents NASM, the Netwide Assembler: an assembler targetting the Intel x86 series of processors, with portable source.}
46 \M{infotitle}{The Netwide Assembler for x86}
47 \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 \IR{solaris x86} Solaris x86
244 \IA{standard section names}{standardized section names}
245 \IR{symbols, exporting from dlls} symbols, exporting from DLLs
246 \IR{symbols, importing from dlls} symbols, importing from DLLs
247 \IR{test subdirectory} \c{test} subdirectory
249 \IR{underscore, in c symbols} underscore, in C symbols
255 \IA{sco unix}{unix, sco}
256 \IR{unix, sco} Unix, SCO
257 \IA{unix source archive}{unix, source archive}
258 \IR{unix, source archive} Unix, source archive
259 \IA{unix system v}{unix, system v}
260 \IR{unix, system v} Unix, System V
261 \IR{unixware} UnixWare
263 \IR{version number of nasm} version number of NASM
264 \IR{visual c++} Visual C++
265 \IR{www page} WWW page
269 \IR{windows 95} Windows 95
270 \IR{windows nt} Windows NT
271 \# \IC{program entry point}{entry point, program}
272 \# \IC{program entry point}{start point, program}
273 \# \IC{MS-DOS device drivers}{device drivers, MS-DOS}
274 \# \IC{16-bit mode, versus 32-bit mode}{32-bit mode, versus 16-bit mode}
275 \# \IC{c symbol names}{symbol names, in C}
278 \C{intro} Introduction
280 \H{whatsnasm} What Is NASM?
282 The Netwide Assembler, NASM, is an 80x86 and x86-64 assembler designed
283 for portability and modularity. It supports a range of object file
284 formats, including Linux and \c{*BSD} \c{a.out}, \c{ELF}, \c{COFF},
285 \c{Mach-O}, Microsoft 16-bit \c{OBJ}, \c{Win32} and \c{Win64}. It will
286 also output plain binary files. Its syntax is designed to be simple
287 and easy to understand, similar to Intel's but less complex. It
288 supports all currently known x86 architectural extensions, and has
289 strong support for macros.
292 \S{yaasm} Why Yet Another Assembler?
294 The Netwide Assembler grew out of an idea on \i\c{comp.lang.asm.x86}
295 (or possibly \i\c{alt.lang.asm} - I forget which), which was
296 essentially that there didn't seem to be a good \e{free} x86-series
297 assembler around, and that maybe someone ought to write one.
299 \b \i\c{a86} is good, but not free, and in particular you don't get any
300 32-bit capability until you pay. It's DOS only, too.
302 \b \i\c{gas} is free, and ports over to DOS and Unix, but it's not
303 very good, since it's designed to be a back end to \i\c{gcc}, which
304 always feeds it correct code. So its error checking is minimal. Also,
305 its syntax is horrible, from the point of view of anyone trying to
306 actually \e{write} anything in it. Plus you can't write 16-bit code in
309 \b \i\c{as86} is specific to Minix and Linux, and (my version at least)
310 doesn't seem to have much (or any) documentation.
312 \b \i\c{MASM} isn't very good, and it's (was) expensive, and it runs only under
315 \b \i\c{TASM} is better, but still strives for MASM compatibility,
316 which means millions of directives and tons of red tape. And its syntax
317 is essentially MASM's, with the contradictions and quirks that
318 entails (although it sorts out some of those by means of Ideal mode.)
319 It's expensive too. And it's DOS-only.
321 So here, for your coding pleasure, is NASM. At present it's
322 still in prototype stage - we don't promise that it can outperform
323 any of these assemblers. But please, \e{please} send us bug reports,
324 fixes, helpful information, and anything else you can get your hands
325 on (and thanks to the many people who've done this already! You all
326 know who you are), and we'll improve it out of all recognition.
330 \S{legal} \i{License} Conditions
332 Please see the file \c{LICENSE}, supplied as part of any NASM
333 distribution archive, for the license conditions under which you may
334 use NASM. NASM is now under the so-called 2-clause BSD license, also
335 known as the simplified BSD license.
337 Copyright 1996-2010 the NASM Authors - All rights reserved.
339 Redistribution and use in source and binary forms, with or without
340 modification, are permitted provided that the following conditions are
343 \b Redistributions of source code must retain the above copyright
344 notice, this list of conditions and the following disclaimer.
346 \b Redistributions in binary form must reproduce the above copyright
347 notice, this list of conditions and the following disclaimer in the
348 documentation and/or other materials provided with the distribution.
350 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
351 CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
352 INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
353 MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
354 DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
355 CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
356 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
357 NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
358 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
359 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
360 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
361 OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
362 EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
365 \H{contact} Contact Information
367 The current version of NASM (since about 0.98.08) is maintained by a
368 team of developers, accessible through the \c{nasm-devel} mailing list
369 (see below for the link).
370 If you want to report a bug, please read \k{bugs} first.
372 NASM has a \i{website} at
373 \W{http://www.nasm.us/}\c{http://www.nasm.us/}. If it's not there,
376 \i{New releases}, \i{release candidates}, and \I{snapshots, daily
377 development}\i{daily development snapshots} of NASM are available from
378 the official web site.
380 Announcements are posted to
381 \W{news:comp.lang.asm.x86}\i\c{comp.lang.asm.x86},
383 \W{http://www.freshmeat.net/}\c{http://www.freshmeat.net/}.
385 If you want information about the current development status, please
386 subscribe to the \i\c{nasm-devel} email list; see link from the
390 \H{install} Installation
392 \S{instdos} \i{Installing} NASM under MS-\i{DOS} or Windows
394 Once you've obtained the appropriate archive for NASM,
395 \i\c{nasm-XXX-dos.zip} or \i\c{nasm-XXX-win32.zip} (where \c{XXX}
396 denotes the version number of NASM contained in the archive), unpack
397 it into its own directory (for example \c{c:\\nasm}).
399 The archive will contain a set of executable files: the NASM
400 executable file \i\c{nasm.exe}, the NDISASM executable file
401 \i\c{ndisasm.exe}, and possibly additional utilities to handle the
404 The only file NASM needs to run is its own executable, so copy
405 \c{nasm.exe} to a directory on your PATH, or alternatively edit
406 \i\c{autoexec.bat} to add the \c{nasm} directory to your
407 \i\c{PATH} (to do that under Windows XP, go to Start > Control Panel >
408 System > Advanced > Environment Variables; these instructions may work
409 under other versions of Windows as well.)
411 That's it - NASM is installed. You don't need the nasm directory
412 to be present to run NASM (unless you've added it to your \c{PATH}),
413 so you can delete it if you need to save space; however, you may
414 want to keep the documentation or test programs.
416 If you've downloaded the \i{DOS source archive}, \i\c{nasm-XXX.zip},
417 the \c{nasm} directory will also contain the full NASM \i{source
418 code}, and a selection of \i{Makefiles} you can (hopefully) use to
419 rebuild your copy of NASM from scratch. See the file \c{INSTALL} in
422 Note that a number of files are generated from other files by Perl
423 scripts. Although the NASM source distribution includes these
424 generated files, you will need to rebuild them (and hence, will need a
425 Perl interpreter) if you change insns.dat, standard.mac or the
426 documentation. It is possible future source distributions may not
427 include these files at all. Ports of \i{Perl} for a variety of
428 platforms, including DOS and Windows, are available from
429 \W{http://www.cpan.org/ports/}\i{www.cpan.org}.
432 \S{instdos} Installing NASM under \i{Unix}
434 Once you've obtained the \i{Unix source archive} for NASM,
435 \i\c{nasm-XXX.tar.gz} (where \c{XXX} denotes the version number of
436 NASM contained in the archive), unpack it into a directory such
437 as \c{/usr/local/src}. The archive, when unpacked, will create its
438 own subdirectory \c{nasm-XXX}.
440 NASM is an \I{Autoconf}\I\c{configure}auto-configuring package: once
441 you've unpacked it, \c{cd} to the directory it's been unpacked into
442 and type \c{./configure}. This shell script will find the best C
443 compiler to use for building NASM and set up \i{Makefiles}
446 Once NASM has auto-configured, you can type \i\c{make} to build the
447 \c{nasm} and \c{ndisasm} binaries, and then \c{make install} to
448 install them in \c{/usr/local/bin} and install the \i{man pages}
449 \i\c{nasm.1} and \i\c{ndisasm.1} in \c{/usr/local/man/man1}.
450 Alternatively, you can give options such as \c{--prefix} to the
451 configure script (see the file \i\c{INSTALL} for more details), or
452 install the programs yourself.
454 NASM also comes with a set of utilities for handling the \c{RDOFF}
455 custom object-file format, which are in the \i\c{rdoff} subdirectory
456 of the NASM archive. You can build these with \c{make rdf} and
457 install them with \c{make rdf_install}, if you want them.
460 \C{running} Running NASM
462 \H{syntax} NASM \i{Command-Line} Syntax
464 To assemble a file, you issue a command of the form
466 \c nasm -f <format> <filename> [-o <output>]
470 \c nasm -f elf myfile.asm
472 will assemble \c{myfile.asm} into an \c{ELF} object file \c{myfile.o}. And
474 \c nasm -f bin myfile.asm -o myfile.com
476 will assemble \c{myfile.asm} into a raw binary file \c{myfile.com}.
478 To produce a listing file, with the hex codes output from NASM
479 displayed on the left of the original sources, use the \c{-l} option
480 to give a listing file name, for example:
482 \c nasm -f coff myfile.asm -l myfile.lst
484 To get further usage instructions from NASM, try typing
488 As \c{-hf}, this will also list the available output file formats, and what they
491 If you use Linux but aren't sure whether your system is \c{a.out}
496 (in the directory in which you put the NASM binary when you
497 installed it). If it says something like
499 \c nasm: ELF 32-bit LSB executable i386 (386 and up) Version 1
501 then your system is \c{ELF}, and you should use the option \c{-f elf}
502 when you want NASM to produce Linux object files. If it says
504 \c nasm: Linux/i386 demand-paged executable (QMAGIC)
506 or something similar, your system is \c{a.out}, and you should use
507 \c{-f aout} instead (Linux \c{a.out} systems have long been obsolete,
508 and are rare these days.)
510 Like Unix compilers and assemblers, NASM is silent unless it
511 goes wrong: you won't see any output at all, unless it gives error
515 \S{opt-o} The \i\c{-o} Option: Specifying the Output File Name
517 NASM will normally choose the name of your output file for you;
518 precisely how it does this is dependent on the object file format.
519 For Microsoft object file formats (\c{obj}, \c{win32} and \c{win64}),
520 it will remove the \c{.asm} \i{extension} (or whatever extension you
521 like to use - NASM doesn't care) from your source file name and
522 substitute \c{.obj}. For Unix object file formats (\c{aout}, \c{as86},
523 \c{coff}, \c{elf32}, \c{elf64}, \c{ieee}, \c{macho32} and \c{macho64})
524 it will substitute \c{.o}. For \c{dbg}, \c{rdf}, \c{ith} and \c{srec},
525 it will use \c{.dbg}, \c{.rdf}, \c{.ith} and \c{.srec}, respectively,
526 and for the \c{bin} format it will simply remove the extension, so
527 that \c{myfile.asm} produces the output file \c{myfile}.
529 If the output file already exists, NASM will overwrite it, unless it
530 has the same name as the input file, in which case it will give a
531 warning and use \i\c{nasm.out} as the output file name instead.
533 For situations in which this behaviour is unacceptable, NASM
534 provides the \c{-o} command-line option, which allows you to specify
535 your desired output file name. You invoke \c{-o} by following it
536 with the name you wish for the output file, either with or without
537 an intervening space. For example:
539 \c nasm -f bin program.asm -o program.com
540 \c nasm -f bin driver.asm -odriver.sys
542 Note that this is a small o, and is different from a capital O , which
543 is used to specify the number of optimisation passes required. See \k{opt-O}.
546 \S{opt-f} The \i\c{-f} Option: Specifying the \i{Output File Format}
548 If you do not supply the \c{-f} option to NASM, it will choose an
549 output file format for you itself. In the distribution versions of
550 NASM, the default is always \i\c{bin}; if you've compiled your own
551 copy of NASM, you can redefine \i\c{OF_DEFAULT} at compile time and
552 choose what you want the default to be.
554 Like \c{-o}, the intervening space between \c{-f} and the output
555 file format is optional; so \c{-f elf} and \c{-felf} are both valid.
557 A complete list of the available output file formats can be given by
558 issuing the command \i\c{nasm -hf}.
561 \S{opt-l} The \i\c{-l} Option: Generating a \i{Listing File}
563 If you supply the \c{-l} option to NASM, followed (with the usual
564 optional space) by a file name, NASM will generate a
565 \i{source-listing file} for you, in which addresses and generated
566 code are listed on the left, and the actual source code, with
567 expansions of multi-line macros (except those which specifically
568 request no expansion in source listings: see \k{nolist}) on the
571 \c nasm -f elf myfile.asm -l myfile.lst
573 If a list file is selected, you may turn off listing for a
574 section of your source with \c{[list -]}, and turn it back on
575 with \c{[list +]}, (the default, obviously). There is no "user
576 form" (without the brackets). This can be used to list only
577 sections of interest, avoiding excessively long listings.
580 \S{opt-M} The \i\c{-M} Option: Generate \i{Makefile Dependencies}
582 This option can be used to generate makefile dependencies on stdout.
583 This can be redirected to a file for further processing. For example:
585 \c nasm -M myfile.asm > myfile.dep
588 \S{opt-MG} The \i\c{-MG} Option: Generate \i{Makefile Dependencies}
590 This option can be used to generate makefile dependencies on stdout.
591 This differs from the \c{-M} option in that if a nonexisting file is
592 encountered, it is assumed to be a generated file and is added to the
593 dependency list without a prefix.
596 \S{opt-MF} The \i\c\{-MF} Option: Set Makefile Dependency File
598 This option can be used with the \c{-M} or \c{-MG} options to send the
599 output to a file, rather than to stdout. For example:
601 \c nasm -M -MF myfile.dep myfile.asm
604 \S{opt-MD} The \i\c{-MD} Option: Assemble and Generate Dependencies
606 The \c{-MD} option acts as the combination of the \c{-M} and \c{-MF}
607 options (i.e. a filename has to be specified.) However, unlike the
608 \c{-M} or \c{-MG} options, \c{-MD} does \e{not} inhibit the normal
609 operation of the assembler. Use this to automatically generate
610 updated dependencies with every assembly session. For example:
612 \c nasm -f elf -o myfile.o -MD myfile.dep myfile.asm
615 \S{opt-MT} The \i\c{-MT} Option: Dependency Target Name
617 The \c{-MT} option can be used to override the default name of the
618 dependency target. This is normally the same as the output filename,
619 specified by the \c{-o} option.
622 \S{opt-MQ} The \i\c{-MQ} Option: Dependency Target Name (Quoted)
624 The \c{-MQ} option acts as the \c{-MT} option, except it tries to
625 quote characters that have special meaning in Makefile syntax. This
626 is not foolproof, as not all characters with special meaning are
630 \S{opt-MP} The \i\c{-MP} Option: Emit phony targets
632 When used with any of the dependency generation options, the \c{-MP}
633 option causes NASM to emit a phony target without dependencies for
634 each header file. This prevents Make from complaining if a header
635 file has been removed.
638 \S{opt-F} The \i\c{-F} Option: Selecting a \i{Debug Information Format}
640 This option is used to select the format of the debug information
641 emitted into the output file, to be used by a debugger (or \e{will}
642 be). Prior to version 2.03.01, the use of this switch did \e{not} enable
643 output of the selected debug info format. Use \c{-g}, see \k{opt-g},
644 to enable output. Versions 2.03.01 and later automatically enable \c{-g}
645 if \c{-F} is specified.
647 A complete list of the available debug file formats for an output
648 format can be seen by issuing the command \c{nasm -f <format> -y}. Not
649 all output formats currently support debugging output. See \k{opt-y}.
651 This should not be confused with the \c{-f dbg} output format option which
652 is not built into NASM by default. For information on how
653 to enable it when building from the sources, see \k{dbgfmt}.
656 \S{opt-g} The \i\c{-g} Option: Enabling \i{Debug Information}.
658 This option can be used to generate debugging information in the specified
659 format. See \k{opt-F}. Using \c{-g} without \c{-F} results in emitting
660 debug info in the default format, if any, for the selected output format.
661 If no debug information is currently implemented in the selected output
662 format, \c{-g} is \e{silently ignored}.
665 \S{opt-X} The \i\c{-X} Option: Selecting an \i{Error Reporting Format}
667 This option can be used to select an error reporting format for any
668 error messages that might be produced by NASM.
670 Currently, two error reporting formats may be selected. They are
671 the \c{-Xvc} option and the \c{-Xgnu} option. The GNU format is
672 the default and looks like this:
674 \c filename.asm:65: error: specific error message
676 where \c{filename.asm} is the name of the source file in which the
677 error was detected, \c{65} is the source file line number on which
678 the error was detected, \c{error} is the severity of the error (this
679 could be \c{warning}), and \c{specific error message} is a more
680 detailed text message which should help pinpoint the exact problem.
682 The other format, specified by \c{-Xvc} is the style used by Microsoft
683 Visual C++ and some other programs. It looks like this:
685 \c filename.asm(65) : error: specific error message
687 where the only difference is that the line number is in parentheses
688 instead of being delimited by colons.
690 See also the \c{Visual C++} output format, \k{win32fmt}.
692 \S{opt-Z} The \i\c{-Z} Option: Send Errors to a File
694 Under \I{DOS}\c{MS-DOS} it can be difficult (though there are ways) to
695 redirect the standard-error output of a program to a file. Since
696 NASM usually produces its warning and \i{error messages} on
697 \i\c{stderr}, this can make it hard to capture the errors if (for
698 example) you want to load them into an editor.
700 NASM therefore provides the \c{-Z} option, taking a filename argument
701 which causes errors to be sent to the specified files rather than
702 standard error. Therefore you can \I{redirecting errors}redirect
703 the errors into a file by typing
705 \c nasm -Z myfile.err -f obj myfile.asm
707 In earlier versions of NASM, this option was called \c{-E}, but it was
708 changed since \c{-E} is an option conventionally used for
709 preprocessing only, with disastrous results. See \k{opt-E}.
711 \S{opt-s} The \i\c{-s} Option: Send Errors to \i\c{stdout}
713 The \c{-s} option redirects \i{error messages} to \c{stdout} rather
714 than \c{stderr}, so it can be redirected under \I{DOS}\c{MS-DOS}. To
715 assemble the file \c{myfile.asm} and pipe its output to the \c{more}
716 program, you can type:
718 \c nasm -s -f obj myfile.asm | more
720 See also the \c{-Z} option, \k{opt-Z}.
723 \S{opt-i} The \i\c{-i}\I\c{-I} Option: Include File Search Directories
725 When NASM sees the \i\c{%include} or \i\c{%pathsearch} directive in a
726 source file (see \k{include}, \k{pathsearch} or \k{incbin}), it will
727 search for the given file not only in the current directory, but also
728 in any directories specified on the command line by the use of the
729 \c{-i} option. Therefore you can include files from a \i{macro
730 library}, for example, by typing
732 \c nasm -ic:\macrolib\ -f obj myfile.asm
734 (As usual, a space between \c{-i} and the path name is allowed, and
737 NASM, in the interests of complete source-code portability, does not
738 understand the file naming conventions of the OS it is running on;
739 the string you provide as an argument to the \c{-i} option will be
740 prepended exactly as written to the name of the include file.
741 Therefore the trailing backslash in the above example is necessary.
742 Under Unix, a trailing forward slash is similarly necessary.
744 (You can use this to your advantage, if you're really \i{perverse},
745 by noting that the option \c{-ifoo} will cause \c{%include "bar.i"}
746 to search for the file \c{foobar.i}...)
748 If you want to define a \e{standard} \i{include search path},
749 similar to \c{/usr/include} on Unix systems, you should place one or
750 more \c{-i} directives in the \c{NASMENV} environment variable (see
753 For Makefile compatibility with many C compilers, this option can also
754 be specified as \c{-I}.
757 \S{opt-p} The \i\c{-p}\I\c{-P} Option: \I{pre-including files}Pre-Include a File
759 \I\c{%include}NASM allows you to specify files to be
760 \e{pre-included} into your source file, by the use of the \c{-p}
763 \c nasm myfile.asm -p myinc.inc
765 is equivalent to running \c{nasm myfile.asm} and placing the
766 directive \c{%include "myinc.inc"} at the start of the file.
768 For consistency with the \c{-I}, \c{-D} and \c{-U} options, this
769 option can also be specified as \c{-P}.
772 \S{opt-d} The \i\c{-d}\I\c{-D} Option: \I{pre-defining macros}Pre-Define a Macro
774 \I\c{%define}Just as the \c{-p} option gives an alternative to placing
775 \c{%include} directives at the start of a source file, the \c{-d}
776 option gives an alternative to placing a \c{%define} directive. You
779 \c nasm myfile.asm -dFOO=100
781 as an alternative to placing the directive
785 at the start of the file. You can miss off the macro value, as well:
786 the option \c{-dFOO} is equivalent to coding \c{%define FOO}. This
787 form of the directive may be useful for selecting \i{assembly-time
788 options} which are then tested using \c{%ifdef}, for example
791 For Makefile compatibility with many C compilers, this option can also
792 be specified as \c{-D}.
795 \S{opt-u} The \i\c{-u}\I\c{-U} Option: \I{Undefining macros}Undefine a Macro
797 \I\c{%undef}The \c{-u} option undefines a macro that would otherwise
798 have been pre-defined, either automatically or by a \c{-p} or \c{-d}
799 option specified earlier on the command lines.
801 For example, the following command line:
803 \c nasm myfile.asm -dFOO=100 -uFOO
805 would result in \c{FOO} \e{not} being a predefined macro in the
806 program. This is useful to override options specified at a different
809 For Makefile compatibility with many C compilers, this option can also
810 be specified as \c{-U}.
813 \S{opt-E} The \i\c{-E}\I{-e} Option: Preprocess Only
815 NASM allows the \i{preprocessor} to be run on its own, up to a
816 point. Using the \c{-E} option (which requires no arguments) will
817 cause NASM to preprocess its input file, expand all the macro
818 references, remove all the comments and preprocessor directives, and
819 print the resulting file on standard output (or save it to a file,
820 if the \c{-o} option is also used).
822 This option cannot be applied to programs which require the
823 preprocessor to evaluate \I{preprocessor expressions}\i{expressions}
824 which depend on the values of symbols: so code such as
826 \c %assign tablesize ($-tablestart)
828 will cause an error in \i{preprocess-only mode}.
830 For compatiblity with older version of NASM, this option can also be
831 written \c{-e}. \c{-E} in older versions of NASM was the equivalent
832 of the current \c{-Z} option, \k{opt-Z}.
834 \S{opt-a} The \i\c{-a} Option: Don't Preprocess At All
836 If NASM is being used as the back end to a compiler, it might be
837 desirable to \I{suppressing preprocessing}suppress preprocessing
838 completely and assume the compiler has already done it, to save time
839 and increase compilation speeds. The \c{-a} option, requiring no
840 argument, instructs NASM to replace its powerful \i{preprocessor}
841 with a \i{stub preprocessor} which does nothing.
844 \S{opt-O} The \i\c{-O} Option: Specifying \i{Multipass Optimization}
846 NASM defaults to not optimizing operands which can fit into a signed byte.
847 This means that if you want the shortest possible object code,
848 you have to enable optimization.
850 Using the \c{-O} option, you can tell NASM to carry out different
851 levels of optimization. The syntax is:
853 \b \c{-O0}: No optimization. All operands take their long forms,
854 if a short form is not specified, except conditional jumps.
855 This is intended to match NASM 0.98 behavior.
857 \b \c{-O1}: Minimal optimization. As above, but immediate operands
858 which will fit in a signed byte are optimized,
859 unless the long form is specified. Conditional jumps default
860 to the long form unless otherwise specified.
862 \b \c{-Ox} (where \c{x} is the actual letter \c{x}): Multipass optimization.
863 Minimize branch offsets and signed immediate bytes,
864 overriding size specification unless the \c{strict} keyword
865 has been used (see \k{strict}). For compatability with earlier
866 releases, the letter \c{x} may also be any number greater than
867 one. This number has no effect on the actual number of passes.
869 The \c{-Ox} mode is recommended for most uses.
871 Note that this is a capital \c{O}, and is different from a small \c{o}, which
872 is used to specify the output file name. See \k{opt-o}.
875 \S{opt-t} The \i\c{-t} Option: Enable TASM Compatibility Mode
877 NASM includes a limited form of compatibility with Borland's \i\c{TASM}.
878 When NASM's \c{-t} option is used, the following changes are made:
880 \b local labels may be prefixed with \c{@@} instead of \c{.}
882 \b size override is supported within brackets. In TASM compatible mode,
883 a size override inside square brackets changes the size of the operand,
884 and not the address type of the operand as it does in NASM syntax. E.g.
885 \c{mov eax,[DWORD val]} is valid syntax in TASM compatibility mode.
886 Note that you lose the ability to override the default address type for
889 \b unprefixed forms of some directives supported (\c{arg}, \c{elif},
890 \c{else}, \c{endif}, \c{if}, \c{ifdef}, \c{ifdifi}, \c{ifndef},
891 \c{include}, \c{local})
893 \S{opt-w} The \i\c{-w} and \i\c{-W} Options: Enable or Disable Assembly \i{Warnings}
895 NASM can observe many conditions during the course of assembly which
896 are worth mentioning to the user, but not a sufficiently severe
897 error to justify NASM refusing to generate an output file. These
898 conditions are reported like errors, but come up with the word
899 `warning' before the message. Warnings do not prevent NASM from
900 generating an output file and returning a success status to the
903 Some conditions are even less severe than that: they are only
904 sometimes worth mentioning to the user. Therefore NASM supports the
905 \c{-w} command-line option, which enables or disables certain
906 classes of assembly warning. Such warning classes are described by a
907 name, for example \c{orphan-labels}; you can enable warnings of
908 this class by the command-line option \c{-w+orphan-labels} and
909 disable it by \c{-w-orphan-labels}.
911 The \i{suppressible warning} classes are:
913 \b \i\c{macro-params} covers warnings about \i{multi-line macros}
914 being invoked with the wrong number of parameters. This warning
915 class is enabled by default; see \k{mlmacover} for an example of why
916 you might want to disable it.
918 \b \i\c{macro-selfref} warns if a macro references itself. This
919 warning class is disabled by default.
921 \b\i\c{macro-defaults} warns when a macro has more default
922 parameters than optional parameters. This warning class
923 is enabled by default; see \k{mlmacdef} for why you might want to disable it.
925 \b \i\c{orphan-labels} covers warnings about source lines which
926 contain no instruction but define a label without a trailing colon.
927 NASM warns about this somewhat obscure condition by default;
928 see \k{syntax} for more information.
930 \b \i\c{number-overflow} covers warnings about numeric constants which
931 don't fit in 64 bits. This warning class is enabled by default.
933 \b \i\c{gnu-elf-extensions} warns if 8-bit or 16-bit relocations
934 are used in \c{-f elf} format. The GNU extensions allow this.
935 This warning class is disabled by default.
937 \b \i\c{float-overflow} warns about floating point overflow.
940 \b \i\c{float-denorm} warns about floating point denormals.
943 \b \i\c{float-underflow} warns about floating point underflow.
946 \b \i\c{float-toolong} warns about too many digits in floating-point numbers.
949 \b \i\c{user} controls \c{%warning} directives (see \k{pperror}).
952 \b \i\c{error} causes warnings to be treated as errors. Disabled by
955 \b \i\c{all} is an alias for \e{all} suppressible warning classes (not
956 including \c{error}). Thus, \c{-w+all} enables all available warnings.
958 In addition, you can set warning classes across sections.
959 Warning classes may be enabled with \i\c{[warning +warning-name]},
960 disabled with \i\c{[warning -warning-name]} or reset to their
961 original value with \i\c{[warning *warning-name]}. No "user form"
962 (without the brackets) exists.
964 Since version 2.00, NASM has also supported the gcc-like syntax
965 \c{-Wwarning} and \c{-Wno-warning} instead of \c{-w+warning} and
966 \c{-w-warning}, respectively.
969 \S{opt-v} The \i\c{-v} Option: Display \i{Version} Info
971 Typing \c{NASM -v} will display the version of NASM which you are using,
972 and the date on which it was compiled.
974 You will need the version number if you report a bug.
976 \S{opt-y} The \i\c{-y} Option: Display Available Debug Info Formats
978 Typing \c{nasm -f <option> -y} will display a list of the available
979 debug info formats for the given output format. The default format
980 is indicated by an asterisk. For example:
984 \c valid debug formats for 'elf32' output format are
985 \c ('*' denotes default):
986 \c * stabs ELF32 (i386) stabs debug format for Linux
987 \c dwarf elf32 (i386) dwarf debug format for Linux
990 \S{opt-pfix} The \i\c{--prefix} and \i\c{--postfix} Options.
992 The \c{--prefix} and \c{--postfix} options prepend or append
993 (respectively) the given argument to all \c{global} or
994 \c{extern} variables. E.g. \c{--prefix _} will prepend the
995 underscore to all global and external variables, as C sometimes
996 (but not always) likes it.
999 \S{nasmenv} The \i\c{NASMENV} \i{Environment} Variable
1001 If you define an environment variable called \c{NASMENV}, the program
1002 will interpret it as a list of extra command-line options, which are
1003 processed before the real command line. You can use this to define
1004 standard search directories for include files, by putting \c{-i}
1005 options in the \c{NASMENV} variable.
1007 The value of the variable is split up at white space, so that the
1008 value \c{-s -ic:\\nasmlib\\} will be treated as two separate options.
1009 However, that means that the value \c{-dNAME="my name"} won't do
1010 what you might want, because it will be split at the space and the
1011 NASM command-line processing will get confused by the two
1012 nonsensical words \c{-dNAME="my} and \c{name"}.
1014 To get round this, NASM provides a feature whereby, if you begin the
1015 \c{NASMENV} environment variable with some character that isn't a minus
1016 sign, then NASM will treat this character as the \i{separator
1017 character} for options. So setting the \c{NASMENV} variable to the
1018 value \c{!-s!-ic:\\nasmlib\\} is equivalent to setting it to \c{-s
1019 -ic:\\nasmlib\\}, but \c{!-dNAME="my name"} will work.
1021 This environment variable was previously called \c{NASM}. This was
1022 changed with version 0.98.31.
1025 \H{qstart} \i{Quick Start} for \i{MASM} Users
1027 If you're used to writing programs with MASM, or with \i{TASM} in
1028 MASM-compatible (non-Ideal) mode, or with \i\c{a86}, this section
1029 attempts to outline the major differences between MASM's syntax and
1030 NASM's. If you're not already used to MASM, it's probably worth
1031 skipping this section.
1034 \S{qscs} NASM Is \I{case sensitivity}Case-Sensitive
1036 One simple difference is that NASM is case-sensitive. It makes a
1037 difference whether you call your label \c{foo}, \c{Foo} or \c{FOO}.
1038 If you're assembling to \c{DOS} or \c{OS/2} \c{.OBJ} files, you can
1039 invoke the \i\c{UPPERCASE} directive (documented in \k{objfmt}) to
1040 ensure that all symbols exported to other code modules are forced
1041 to be upper case; but even then, \e{within} a single module, NASM
1042 will distinguish between labels differing only in case.
1045 \S{qsbrackets} NASM Requires \i{Square Brackets} For \i{Memory References}
1047 NASM was designed with simplicity of syntax in mind. One of the
1048 \i{design goals} of NASM is that it should be possible, as far as is
1049 practical, for the user to look at a single line of NASM code
1050 and tell what opcode is generated by it. You can't do this in MASM:
1051 if you declare, for example,
1056 then the two lines of code
1061 generate completely different opcodes, despite having
1062 identical-looking syntaxes.
1064 NASM avoids this undesirable situation by having a much simpler
1065 syntax for memory references. The rule is simply that any access to
1066 the \e{contents} of a memory location requires square brackets
1067 around the address, and any access to the \e{address} of a variable
1068 doesn't. So an instruction of the form \c{mov ax,foo} will
1069 \e{always} refer to a compile-time constant, whether it's an \c{EQU}
1070 or the address of a variable; and to access the \e{contents} of the
1071 variable \c{bar}, you must code \c{mov ax,[bar]}.
1073 This also means that NASM has no need for MASM's \i\c{OFFSET}
1074 keyword, since the MASM code \c{mov ax,offset bar} means exactly the
1075 same thing as NASM's \c{mov ax,bar}. If you're trying to get
1076 large amounts of MASM code to assemble sensibly under NASM, you
1077 can always code \c{%idefine offset} to make the preprocessor treat
1078 the \c{OFFSET} keyword as a no-op.
1080 This issue is even more confusing in \i\c{a86}, where declaring a
1081 label with a trailing colon defines it to be a `label' as opposed to
1082 a `variable' and causes \c{a86} to adopt NASM-style semantics; so in
1083 \c{a86}, \c{mov ax,var} has different behaviour depending on whether
1084 \c{var} was declared as \c{var: dw 0} (a label) or \c{var dw 0} (a
1085 word-size variable). NASM is very simple by comparison:
1086 \e{everything} is a label.
1088 NASM, in the interests of simplicity, also does not support the
1089 \i{hybrid syntaxes} supported by MASM and its clones, such as
1090 \c{mov ax,table[bx]}, where a memory reference is denoted by one
1091 portion outside square brackets and another portion inside. The
1092 correct syntax for the above is \c{mov ax,[table+bx]}. Likewise,
1093 \c{mov ax,es:[di]} is wrong and \c{mov ax,[es:di]} is right.
1096 \S{qstypes} NASM Doesn't Store \i{Variable Types}
1098 NASM, by design, chooses not to remember the types of variables you
1099 declare. Whereas MASM will remember, on seeing \c{var dw 0}, that
1100 you declared \c{var} as a word-size variable, and will then be able
1101 to fill in the \i{ambiguity} in the size of the instruction \c{mov
1102 var,2}, NASM will deliberately remember nothing about the symbol
1103 \c{var} except where it begins, and so you must explicitly code
1104 \c{mov word [var],2}.
1106 For this reason, NASM doesn't support the \c{LODS}, \c{MOVS},
1107 \c{STOS}, \c{SCAS}, \c{CMPS}, \c{INS}, or \c{OUTS} instructions,
1108 but only supports the forms such as \c{LODSB}, \c{MOVSW}, and
1109 \c{SCASD}, which explicitly specify the size of the components of
1110 the strings being manipulated.
1113 \S{qsassume} NASM Doesn't \i\c{ASSUME}
1115 As part of NASM's drive for simplicity, it also does not support the
1116 \c{ASSUME} directive. NASM will not keep track of what values you
1117 choose to put in your segment registers, and will never
1118 \e{automatically} generate a \i{segment override} prefix.
1121 \S{qsmodel} NASM Doesn't Support \i{Memory Models}
1123 NASM also does not have any directives to support different 16-bit
1124 memory models. The programmer has to keep track of which functions
1125 are supposed to be called with a \i{far call} and which with a
1126 \i{near call}, and is responsible for putting the correct form of
1127 \c{RET} instruction (\c{RETN} or \c{RETF}; NASM accepts \c{RET}
1128 itself as an alternate form for \c{RETN}); in addition, the
1129 programmer is responsible for coding CALL FAR instructions where
1130 necessary when calling \e{external} functions, and must also keep
1131 track of which external variable definitions are far and which are
1135 \S{qsfpu} \i{Floating-Point} Differences
1137 NASM uses different names to refer to floating-point registers from
1138 MASM: where MASM would call them \c{ST(0)}, \c{ST(1)} and so on, and
1139 \i\c{a86} would call them simply \c{0}, \c{1} and so on, NASM
1140 chooses to call them \c{st0}, \c{st1} etc.
1142 As of version 0.96, NASM now treats the instructions with
1143 \i{`nowait'} forms in the same way as MASM-compatible assemblers.
1144 The idiosyncratic treatment employed by 0.95 and earlier was based
1145 on a misunderstanding by the authors.
1148 \S{qsother} Other Differences
1150 For historical reasons, NASM uses the keyword \i\c{TWORD} where MASM
1151 and compatible assemblers use \i\c{TBYTE}.
1153 NASM does not declare \i{uninitialized storage} in the same way as
1154 MASM: where a MASM programmer might use \c{stack db 64 dup (?)},
1155 NASM requires \c{stack resb 64}, intended to be read as `reserve 64
1156 bytes'. For a limited amount of compatibility, since NASM treats
1157 \c{?} as a valid character in symbol names, you can code \c{? equ 0}
1158 and then writing \c{dw ?} will at least do something vaguely useful.
1159 \I\c{RESB}\i\c{DUP} is still not a supported syntax, however.
1161 In addition to all of this, macros and directives work completely
1162 differently to MASM. See \k{preproc} and \k{directive} for further
1166 \C{lang} The NASM Language
1168 \H{syntax} Layout of a NASM Source Line
1170 Like most assemblers, each NASM source line contains (unless it
1171 is a macro, a preprocessor directive or an assembler directive: see
1172 \k{preproc} and \k{directive}) some combination of the four fields
1174 \c label: instruction operands ; comment
1176 As usual, most of these fields are optional; the presence or absence
1177 of any combination of a label, an instruction and a comment is allowed.
1178 Of course, the operand field is either required or forbidden by the
1179 presence and nature of the instruction field.
1181 NASM uses backslash (\\) as the line continuation character; if a line
1182 ends with backslash, the next line is considered to be a part of the
1183 backslash-ended line.
1185 NASM places no restrictions on white space within a line: labels may
1186 have white space before them, or instructions may have no space
1187 before them, or anything. The \i{colon} after a label is also
1188 optional. (Note that this means that if you intend to code \c{lodsb}
1189 alone on a line, and type \c{lodab} by accident, then that's still a
1190 valid source line which does nothing but define a label. Running
1191 NASM with the command-line option
1192 \I{orphan-labels}\c{-w+orphan-labels} will cause it to warn you if
1193 you define a label alone on a line without a \i{trailing colon}.)
1195 \i{Valid characters} in labels are letters, numbers, \c{_}, \c{$},
1196 \c{#}, \c{@}, \c{~}, \c{.}, and \c{?}. The only characters which may
1197 be used as the \e{first} character of an identifier are letters,
1198 \c{.} (with special meaning: see \k{locallab}), \c{_} and \c{?}.
1199 An identifier may also be prefixed with a \I{$, prefix}\c{$} to
1200 indicate that it is intended to be read as an identifier and not a
1201 reserved word; thus, if some other module you are linking with
1202 defines a symbol called \c{eax}, you can refer to \c{$eax} in NASM
1203 code to distinguish the symbol from the register. Maximum length of
1204 an identifier is 4095 characters.
1206 The instruction field may contain any machine instruction: Pentium
1207 and P6 instructions, FPU instructions, MMX instructions and even
1208 undocumented instructions are all supported. The instruction may be
1209 prefixed by \c{LOCK}, \c{REP}, \c{REPE}/\c{REPZ} or
1210 \c{REPNE}/\c{REPNZ}, in the usual way. Explicit \I{address-size
1211 prefixes}address-size and \i{operand-size prefixes} \i\c{A16},
1212 \i\c{A32}, \i\c{A64}, \i\c{O16} and \i\c{O32}, \i\c{O64} are provided - one example of their use
1213 is given in \k{mixsize}. You can also use the name of a \I{segment
1214 override}segment register as an instruction prefix: coding
1215 \c{es mov [bx],ax} is equivalent to coding \c{mov [es:bx],ax}. We
1216 recommend the latter syntax, since it is consistent with other
1217 syntactic features of the language, but for instructions such as
1218 \c{LODSB}, which has no operands and yet can require a segment
1219 override, there is no clean syntactic way to proceed apart from
1222 An instruction is not required to use a prefix: prefixes such as
1223 \c{CS}, \c{A32}, \c{LOCK} or \c{REPE} can appear on a line by
1224 themselves, and NASM will just generate the prefix bytes.
1226 In addition to actual machine instructions, NASM also supports a
1227 number of pseudo-instructions, described in \k{pseudop}.
1229 Instruction \i{operands} may take a number of forms: they can be
1230 registers, described simply by the register name (e.g. \c{ax},
1231 \c{bp}, \c{ebx}, \c{cr0}: NASM does not use the \c{gas}-style
1232 syntax in which register names must be prefixed by a \c{%} sign), or
1233 they can be \i{effective addresses} (see \k{effaddr}), constants
1234 (\k{const}) or expressions (\k{expr}).
1236 For x87 \i{floating-point} instructions, NASM accepts a wide range of
1237 syntaxes: you can use two-operand forms like MASM supports, or you
1238 can use NASM's native single-operand forms in most cases.
1240 \# all forms of each supported instruction are given in
1242 For example, you can code:
1244 \c fadd st1 ; this sets st0 := st0 + st1
1245 \c fadd st0,st1 ; so does this
1247 \c fadd st1,st0 ; this sets st1 := st1 + st0
1248 \c fadd to st1 ; so does this
1250 Almost any x87 floating-point instruction that references memory must
1251 use one of the prefixes \i\c{DWORD}, \i\c{QWORD} or \i\c{TWORD} to
1252 indicate what size of \i{memory operand} it refers to.
1255 \H{pseudop} \i{Pseudo-Instructions}
1257 Pseudo-instructions are things which, though not real x86 machine
1258 instructions, are used in the instruction field anyway because that's
1259 the most convenient place to put them. The current pseudo-instructions
1260 are \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, \i\c{DO} and
1261 \i\c{DY}; their \i{uninitialized} counterparts \i\c{RESB}, \i\c{RESW},
1262 \i\c{RESD}, \i\c{RESQ}, \i\c{REST}, \i\c{RESO} and \i\c{RESY}; the
1263 \i\c{INCBIN} command, the \i\c{EQU} command, and the \i\c{TIMES}
1267 \S{db} \c{DB} and Friends: Declaring Initialized Data
1269 \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, \i\c{DO} and
1270 \i\c{DY} are used, much as in MASM, to declare initialized data in the
1271 output file. They can be invoked in a wide range of ways:
1272 \I{floating-point}\I{character constant}\I{string constant}
1274 \c db 0x55 ; just the byte 0x55
1275 \c db 0x55,0x56,0x57 ; three bytes in succession
1276 \c db 'a',0x55 ; character constants are OK
1277 \c db 'hello',13,10,'$' ; so are string constants
1278 \c dw 0x1234 ; 0x34 0x12
1279 \c dw 'a' ; 0x61 0x00 (it's just a number)
1280 \c dw 'ab' ; 0x61 0x62 (character constant)
1281 \c dw 'abc' ; 0x61 0x62 0x63 0x00 (string)
1282 \c dd 0x12345678 ; 0x78 0x56 0x34 0x12
1283 \c dd 1.234567e20 ; floating-point constant
1284 \c dq 0x123456789abcdef0 ; eight byte constant
1285 \c dq 1.234567e20 ; double-precision float
1286 \c dt 1.234567e20 ; extended-precision float
1288 \c{DT}, \c{DO} and \c{DY} do not accept \i{numeric constants} as operands.
1291 \S{resb} \c{RESB} and Friends: Declaring \i{Uninitialized} Data
1293 \i\c{RESB}, \i\c{RESW}, \i\c{RESD}, \i\c{RESQ}, \i\c{REST}, \i\c{RESO}
1294 and \i\c{RESY} are designed to be used in the BSS section of a module:
1295 they declare \e{uninitialized} storage space. Each takes a single
1296 operand, which is the number of bytes, words, doublewords or whatever
1297 to reserve. As stated in \k{qsother}, NASM does not support the
1298 MASM/TASM syntax of reserving uninitialized space by writing
1299 \I\c{?}\c{DW ?} or similar things: this is what it does instead. The
1300 operand to a \c{RESB}-type pseudo-instruction is a \i\e{critical
1301 expression}: see \k{crit}.
1305 \c buffer: resb 64 ; reserve 64 bytes
1306 \c wordvar: resw 1 ; reserve a word
1307 \c realarray resq 10 ; array of ten reals
1308 \c ymmval: resy 1 ; one YMM register
1310 \S{incbin} \i\c{INCBIN}: Including External \i{Binary Files}
1312 \c{INCBIN} is borrowed from the old Amiga assembler \i{DevPac}: it
1313 includes a binary file verbatim into the output file. This can be
1314 handy for (for example) including \i{graphics} and \i{sound} data
1315 directly into a game executable file. It can be called in one of
1318 \c incbin "file.dat" ; include the whole file
1319 \c incbin "file.dat",1024 ; skip the first 1024 bytes
1320 \c incbin "file.dat",1024,512 ; skip the first 1024, and
1321 \c ; actually include at most 512
1323 \c{INCBIN} is both a directive and a standard macro; the standard
1324 macro version searches for the file in the include file search path
1325 and adds the file to the dependency lists. This macro can be
1326 overridden if desired.
1329 \S{equ} \i\c{EQU}: Defining Constants
1331 \c{EQU} defines a symbol to a given constant value: when \c{EQU} is
1332 used, the source line must contain a label. The action of \c{EQU} is
1333 to define the given label name to the value of its (only) operand.
1334 This definition is absolute, and cannot change later. So, for
1337 \c message db 'hello, world'
1338 \c msglen equ $-message
1340 defines \c{msglen} to be the constant 12. \c{msglen} may not then be
1341 redefined later. This is not a \i{preprocessor} definition either:
1342 the value of \c{msglen} is evaluated \e{once}, using the value of
1343 \c{$} (see \k{expr} for an explanation of \c{$}) at the point of
1344 definition, rather than being evaluated wherever it is referenced
1345 and using the value of \c{$} at the point of reference.
1348 \S{times} \i\c{TIMES}: \i{Repeating} Instructions or Data
1350 The \c{TIMES} prefix causes the instruction to be assembled multiple
1351 times. This is partly present as NASM's equivalent of the \i\c{DUP}
1352 syntax supported by \i{MASM}-compatible assemblers, in that you can
1355 \c zerobuf: times 64 db 0
1357 or similar things; but \c{TIMES} is more versatile than that. The
1358 argument to \c{TIMES} is not just a numeric constant, but a numeric
1359 \e{expression}, so you can do things like
1361 \c buffer: db 'hello, world'
1362 \c times 64-$+buffer db ' '
1364 which will store exactly enough spaces to make the total length of
1365 \c{buffer} up to 64. Finally, \c{TIMES} can be applied to ordinary
1366 instructions, so you can code trivial \i{unrolled loops} in it:
1370 Note that there is no effective difference between \c{times 100 resb
1371 1} and \c{resb 100}, except that the latter will be assembled about
1372 100 times faster due to the internal structure of the assembler.
1374 The operand to \c{TIMES} is a critical expression (\k{crit}).
1376 Note also that \c{TIMES} can't be applied to \i{macros}: the reason
1377 for this is that \c{TIMES} is processed after the macro phase, which
1378 allows the argument to \c{TIMES} to contain expressions such as
1379 \c{64-$+buffer} as above. To repeat more than one line of code, or a
1380 complex macro, use the preprocessor \i\c{%rep} directive.
1383 \H{effaddr} Effective Addresses
1385 An \i{effective address} is any operand to an instruction which
1386 \I{memory reference}references memory. Effective addresses, in NASM,
1387 have a very simple syntax: they consist of an expression evaluating
1388 to the desired address, enclosed in \i{square brackets}. For
1393 \c mov ax,[wordvar+1]
1394 \c mov ax,[es:wordvar+bx]
1396 Anything not conforming to this simple system is not a valid memory
1397 reference in NASM, for example \c{es:wordvar[bx]}.
1399 More complicated effective addresses, such as those involving more
1400 than one register, work in exactly the same way:
1402 \c mov eax,[ebx*2+ecx+offset]
1405 NASM is capable of doing \i{algebra} on these effective addresses,
1406 so that things which don't necessarily \e{look} legal are perfectly
1409 \c mov eax,[ebx*5] ; assembles as [ebx*4+ebx]
1410 \c mov eax,[label1*2-label2] ; ie [label1+(label1-label2)]
1412 Some forms of effective address have more than one assembled form;
1413 in most such cases NASM will generate the smallest form it can. For
1414 example, there are distinct assembled forms for the 32-bit effective
1415 addresses \c{[eax*2+0]} and \c{[eax+eax]}, and NASM will generally
1416 generate the latter on the grounds that the former requires four
1417 bytes to store a zero offset.
1419 NASM has a hinting mechanism which will cause \c{[eax+ebx]} and
1420 \c{[ebx+eax]} to generate different opcodes; this is occasionally
1421 useful because \c{[esi+ebp]} and \c{[ebp+esi]} have different
1422 default segment registers.
1424 However, you can force NASM to generate an effective address in a
1425 particular form by the use of the keywords \c{BYTE}, \c{WORD},
1426 \c{DWORD} and \c{NOSPLIT}. If you need \c{[eax+3]} to be assembled
1427 using a double-word offset field instead of the one byte NASM will
1428 normally generate, you can code \c{[dword eax+3]}. Similarly, you
1429 can force NASM to use a byte offset for a small value which it
1430 hasn't seen on the first pass (see \k{crit} for an example of such a
1431 code fragment) by using \c{[byte eax+offset]}. As special cases,
1432 \c{[byte eax]} will code \c{[eax+0]} with a byte offset of zero, and
1433 \c{[dword eax]} will code it with a double-word offset of zero. The
1434 normal form, \c{[eax]}, will be coded with no offset field.
1436 The form described in the previous paragraph is also useful if you
1437 are trying to access data in a 32-bit segment from within 16 bit code.
1438 For more information on this see the section on mixed-size addressing
1439 (\k{mixaddr}). In particular, if you need to access data with a known
1440 offset that is larger than will fit in a 16-bit value, if you don't
1441 specify that it is a dword offset, nasm will cause the high word of
1442 the offset to be lost.
1444 Similarly, NASM will split \c{[eax*2]} into \c{[eax+eax]} because
1445 that allows the offset field to be absent and space to be saved; in
1446 fact, it will also split \c{[eax*2+offset]} into
1447 \c{[eax+eax+offset]}. You can combat this behaviour by the use of
1448 the \c{NOSPLIT} keyword: \c{[nosplit eax*2]} will force
1449 \c{[eax*2+0]} to be generated literally.
1451 In 64-bit mode, NASM will by default generate absolute addresses. The
1452 \i\c{REL} keyword makes it produce \c{RIP}-relative addresses. Since
1453 this is frequently the normally desired behaviour, see the \c{DEFAULT}
1454 directive (\k{default}). The keyword \i\c{ABS} overrides \i\c{REL}.
1457 \H{const} \i{Constants}
1459 NASM understands four different types of constant: numeric,
1460 character, string and floating-point.
1463 \S{numconst} \i{Numeric Constants}
1465 A numeric constant is simply a number. NASM allows you to specify
1466 numbers in a variety of number bases, in a variety of ways: you can
1467 suffix \c{H} or \c{X}, \c{Q} or \c{O}, and \c{B} for \i{hexadecimal},
1468 \i{octal} and \i{binary} respectively, or you can prefix \c{0x} for
1469 hexadecimal in the style of C, or you can prefix \c{$} for hexadecimal
1470 in the style of Borland Pascal. Note, though, that the \I{$,
1471 prefix}\c{$} prefix does double duty as a prefix on identifiers (see
1472 \k{syntax}), so a hex number prefixed with a \c{$} sign must have a
1473 digit after the \c{$} rather than a letter. In addition, current
1474 versions of NASM accept the prefix \c{0h} for hexadecimal, \c{0o} or
1475 \c{0q} for octal, and \c{0b} for binary. Please note that unlike C, a
1476 \c{0} prefix by itself does \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 ; hex yet again
1495 \c mov ax,11001000b ; binary
1496 \c mov ax,1100_1000b ; same binary constant
1497 \c mov ax,0b1100_1000 ; same binary constant yet again
1499 \S{strings} \I{Strings}\i{Character Strings}
1501 A character string consists of up to eight characters enclosed in
1502 either single quotes (\c{'...'}), double quotes (\c{"..."}) or
1503 backquotes (\c{`...`}). Single or double quotes are equivalent to
1504 NASM (except of course that surrounding the constant with single
1505 quotes allows double quotes to appear within it and vice versa); the
1506 contents of those are represented verbatim. Strings enclosed in
1507 backquotes support C-style \c{\\}-escapes for special characters.
1510 The following \i{escape sequences} are recognized by backquoted strings:
1512 \c \' single quote (')
1513 \c \" double quote (")
1515 \c \\\ backslash (\)
1516 \c \? question mark (?)
1524 \c \e ESC (ASCII 27)
1525 \c \377 Up to 3 octal digits - literal byte
1526 \c \xFF Up to 2 hexadecimal digits - literal byte
1527 \c \u1234 4 hexadecimal digits - Unicode character
1528 \c \U12345678 8 hexadecimal digits - Unicode character
1530 All other escape sequences are reserved. Note that \c{\\0}, meaning a
1531 \c{NUL} character (ASCII 0), is a special case of the octal escape
1534 \i{Unicode} characters specified with \c{\\u} or \c{\\U} are converted to
1535 \i{UTF-8}. For example, the following lines are all equivalent:
1537 \c db `\u263a` ; UTF-8 smiley face
1538 \c db `\xe2\x98\xba` ; UTF-8 smiley face
1539 \c db 0E2h, 098h, 0BAh ; UTF-8 smiley face
1542 \S{chrconst} \i{Character Constants}
1544 A character constant consists of a string up to eight bytes long, used
1545 in an expression context. It is treated as if it was an integer.
1547 A character constant with more than one byte will be arranged
1548 with \i{little-endian} order in mind: if you code
1552 then the constant generated is not \c{0x61626364}, but
1553 \c{0x64636261}, so that if you were then to store the value into
1554 memory, it would read \c{abcd} rather than \c{dcba}. This is also
1555 the sense of character constants understood by the Pentium's
1556 \i\c{CPUID} instruction.
1559 \S{strconst} \i{String Constants}
1561 String constants are character strings used in the context of some
1562 pseudo-instructions, namely the
1563 \I\c{DW}\I\c{DD}\I\c{DQ}\I\c{DT}\I\c{DO}\I\c{DY}\i\c{DB} family and
1564 \i\c{INCBIN} (where it represents a filename.) They are also used in
1565 certain preprocessor directives.
1567 A string constant looks like a character constant, only longer. It
1568 is treated as a concatenation of maximum-size character constants
1569 for the conditions. So the following are equivalent:
1571 \c db 'hello' ; string constant
1572 \c db 'h','e','l','l','o' ; equivalent character constants
1574 And the following are also equivalent:
1576 \c dd 'ninechars' ; doubleword string constant
1577 \c dd 'nine','char','s' ; becomes three doublewords
1578 \c db 'ninechars',0,0,0 ; and really looks like this
1580 Note that when used in a string-supporting context, quoted strings are
1581 treated as a string constants even if they are short enough to be a
1582 character constant, because otherwise \c{db 'ab'} would have the same
1583 effect as \c{db 'a'}, which would be silly. Similarly, three-character
1584 or four-character constants are treated as strings when they are
1585 operands to \c{DW}, and so forth.
1587 \S{unicode} \I{UTF-16}\I{UTF-32}\i{Unicode} Strings
1589 The special operators \i\c{__utf16__} and \i\c{__utf32__} allows
1590 definition of Unicode strings. They take a string in UTF-8 format and
1591 converts it to (littleendian) UTF-16 or UTF-32, respectively.
1595 \c %define u(x) __utf16__(x)
1596 \c %define w(x) __utf32__(x)
1598 \c dw u('C:\WINDOWS'), 0 ; Pathname in UTF-16
1599 \c dd w(`A + B = \u206a`), 0 ; String in UTF-32
1601 \c{__utf16__} and \c{__utf32__} can be applied either to strings
1602 passed to the \c{DB} family instructions, or to character constants in
1603 an expression context.
1605 \S{fltconst} \I{floating-point, constants}Floating-Point Constants
1607 \i{Floating-point} constants are acceptable only as arguments to
1608 \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, and \i\c{DO}, or as
1609 arguments to the special operators \i\c{__float8__},
1610 \i\c{__float16__}, \i\c{__float32__}, \i\c{__float64__},
1611 \i\c{__float80m__}, \i\c{__float80e__}, \i\c{__float128l__}, and
1612 \i\c{__float128h__}.
1614 Floating-point constants are expressed in the traditional form:
1615 digits, then a period, then optionally more digits, then optionally an
1616 \c{E} followed by an exponent. The period is mandatory, so that NASM
1617 can distinguish between \c{dd 1}, which declares an integer constant,
1618 and \c{dd 1.0} which declares a floating-point constant. NASM also
1619 support C99-style hexadecimal floating-point: \c{0x}, hexadecimal
1620 digits, period, optionally more hexadeximal digits, then optionally a
1621 \c{P} followed by a \e{binary} (not hexadecimal) exponent in decimal
1624 Underscores to break up groups of digits are permitted in
1625 floating-point constants as well.
1629 \c db -0.2 ; "Quarter precision"
1630 \c dw -0.5 ; IEEE 754r/SSE5 half precision
1631 \c dd 1.2 ; an easy one
1632 \c dd 1.222_222_222 ; underscores are permitted
1633 \c dd 0x1p+2 ; 1.0x2^2 = 4.0
1634 \c dq 0x1p+32 ; 1.0x2^32 = 4 294 967 296.0
1635 \c dq 1.e10 ; 10 000 000 000.0
1636 \c dq 1.e+10 ; synonymous with 1.e10
1637 \c dq 1.e-10 ; 0.000 000 000 1
1638 \c dt 3.141592653589793238462 ; pi
1639 \c do 1.e+4000 ; IEEE 754r quad precision
1641 The 8-bit "quarter-precision" floating-point format is
1642 sign:exponent:mantissa = 1:4:3 with an exponent bias of 7. This
1643 appears to be the most frequently used 8-bit floating-point format,
1644 although it is not covered by any formal standard. This is sometimes
1645 called a "\i{minifloat}."
1647 The special operators are used to produce floating-point numbers in
1648 other contexts. They produce the binary representation of a specific
1649 floating-point number as an integer, and can use anywhere integer
1650 constants are used in an expression. \c{__float80m__} and
1651 \c{__float80e__} produce the 64-bit mantissa and 16-bit exponent of an
1652 80-bit floating-point number, and \c{__float128l__} and
1653 \c{__float128h__} produce the lower and upper 64-bit halves of a 128-bit
1654 floating-point number, respectively.
1658 \c mov rax,__float64__(3.141592653589793238462)
1660 ... would assign the binary representation of pi as a 64-bit floating
1661 point number into \c{RAX}. This is exactly equivalent to:
1663 \c mov rax,0x400921fb54442d18
1665 NASM cannot do compile-time arithmetic on floating-point constants.
1666 This is because NASM is designed to be portable - although it always
1667 generates code to run on x86 processors, the assembler itself can
1668 run on any system with an ANSI C compiler. Therefore, the assembler
1669 cannot guarantee the presence of a floating-point unit capable of
1670 handling the \i{Intel number formats}, and so for NASM to be able to
1671 do floating arithmetic it would have to include its own complete set
1672 of floating-point routines, which would significantly increase the
1673 size of the assembler for very little benefit.
1675 The special tokens \i\c{__Infinity__}, \i\c{__QNaN__} (or
1676 \i\c{__NaN__}) and \i\c{__SNaN__} can be used to generate
1677 \I{infinity}infinities, quiet \i{NaN}s, and signalling NaNs,
1678 respectively. These are normally used as macros:
1680 \c %define Inf __Infinity__
1681 \c %define NaN __QNaN__
1683 \c dq +1.5, -Inf, NaN ; Double-precision constants
1685 \S{bcdconst} \I{floating-point, packed BCD constants}Packed BCD Constants
1687 x87-style packed BCD constants can be used in the same contexts as
1688 80-bit floating-point numbers. They are suffixed with \c{p} or
1689 prefixed with \c{0p}, and can include up to 18 decimal digits.
1691 As with other numeric constants, underscores can be used to separate
1696 \c dt 12_345_678_901_245_678p
1697 \c dt -12_345_678_901_245_678p
1702 \H{expr} \i{Expressions}
1704 Expressions in NASM are similar in syntax to those in C. Expressions
1705 are evaluated as 64-bit integers which are then adjusted to the
1708 NASM supports two special tokens in expressions, allowing
1709 calculations to involve the current assembly position: the
1710 \I{$, here}\c{$} and \i\c{$$} tokens. \c{$} evaluates to the assembly
1711 position at the beginning of the line containing the expression; so
1712 you can code an \i{infinite loop} using \c{JMP $}. \c{$$} evaluates
1713 to the beginning of the current section; so you can tell how far
1714 into the section you are by using \c{($-$$)}.
1716 The arithmetic \i{operators} provided by NASM are listed here, in
1717 increasing order of \i{precedence}.
1720 \S{expor} \i\c{|}: \i{Bitwise OR} Operator
1722 The \c{|} operator gives a bitwise OR, exactly as performed by the
1723 \c{OR} machine instruction. Bitwise OR is the lowest-priority
1724 arithmetic operator supported by NASM.
1727 \S{expxor} \i\c{^}: \i{Bitwise XOR} Operator
1729 \c{^} provides the bitwise XOR operation.
1732 \S{expand} \i\c{&}: \i{Bitwise AND} Operator
1734 \c{&} provides the bitwise AND operation.
1737 \S{expshift} \i\c{<<} and \i\c{>>}: \i{Bit Shift} Operators
1739 \c{<<} gives a bit-shift to the left, just as it does in C. So \c{5<<3}
1740 evaluates to 5 times 8, or 40. \c{>>} gives a bit-shift to the
1741 right; in NASM, such a shift is \e{always} unsigned, so that
1742 the bits shifted in from the left-hand end are filled with zero
1743 rather than a sign-extension of the previous highest bit.
1746 \S{expplmi} \I{+ opaddition}\c{+} and \I{- opsubtraction}\c{-}:
1747 \i{Addition} and \i{Subtraction} Operators
1749 The \c{+} and \c{-} operators do perfectly ordinary addition and
1753 \S{expmul} \i\c{*}, \i\c{/}, \i\c{//}, \i\c{%} and \i\c{%%}:
1754 \i{Multiplication} and \i{Division}
1756 \c{*} is the multiplication operator. \c{/} and \c{//} are both
1757 division operators: \c{/} is \i{unsigned division} and \c{//} is
1758 \i{signed division}. Similarly, \c{%} and \c{%%} provide \I{unsigned
1759 modulo}\I{modulo operators}unsigned and
1760 \i{signed modulo} operators respectively.
1762 NASM, like ANSI C, provides no guarantees about the sensible
1763 operation of the signed modulo operator.
1765 Since the \c{%} character is used extensively by the macro
1766 \i{preprocessor}, you should ensure that both the signed and unsigned
1767 modulo operators are followed by white space wherever they appear.
1770 \S{expmul} \i{Unary Operators}: \I{+ opunary}\c{+}, \I{- opunary}\c{-},
1771 \i\c{~}, \I{! opunary}\c{!} and \i\c{SEG}
1773 The highest-priority operators in NASM's expression grammar are
1774 those which only apply to one argument. \c{-} negates its operand,
1775 \c{+} does nothing (it's provided for symmetry with \c{-}), \c{~}
1776 computes the \i{one's complement} of its operand, \c{!} is the
1777 \i{logical negation} operator, and \c{SEG} provides the \i{segment address}
1778 of its operand (explained in more detail in \k{segwrt}).
1781 \H{segwrt} \i\c{SEG} and \i\c{WRT}
1783 When writing large 16-bit programs, which must be split into
1784 multiple \i{segments}, it is often necessary to be able to refer to
1785 the \I{segment address}segment part of the address of a symbol. NASM
1786 supports the \c{SEG} operator to perform this function.
1788 The \c{SEG} operator returns the \i\e{preferred} segment base of a
1789 symbol, defined as the segment base relative to which the offset of
1790 the symbol makes sense. So the code
1792 \c mov ax,seg symbol
1796 will load \c{ES:BX} with a valid pointer to the symbol \c{symbol}.
1798 Things can be more complex than this: since 16-bit segments and
1799 \i{groups} may \I{overlapping segments}overlap, you might occasionally
1800 want to refer to some symbol using a different segment base from the
1801 preferred one. NASM lets you do this, by the use of the \c{WRT}
1802 (With Reference To) keyword. So you can do things like
1804 \c mov ax,weird_seg ; weird_seg is a segment base
1806 \c mov bx,symbol wrt weird_seg
1808 to load \c{ES:BX} with a different, but functionally equivalent,
1809 pointer to the symbol \c{symbol}.
1811 NASM supports far (inter-segment) calls and jumps by means of the
1812 syntax \c{call segment:offset}, where \c{segment} and \c{offset}
1813 both represent immediate values. So to call a far procedure, you
1814 could code either of
1816 \c call (seg procedure):procedure
1817 \c call weird_seg:(procedure wrt weird_seg)
1819 (The parentheses are included for clarity, to show the intended
1820 parsing of the above instructions. They are not necessary in
1823 NASM supports the syntax \I\c{CALL FAR}\c{call far procedure} as a
1824 synonym for the first of the above usages. \c{JMP} works identically
1825 to \c{CALL} in these examples.
1827 To declare a \i{far pointer} to a data item in a data segment, you
1830 \c dw symbol, seg symbol
1832 NASM supports no convenient synonym for this, though you can always
1833 invent one using the macro processor.
1836 \H{strict} \i\c{STRICT}: Inhibiting Optimization
1838 When assembling with the optimizer set to level 2 or higher (see
1839 \k{opt-O}), NASM will use size specifiers (\c{BYTE}, \c{WORD},
1840 \c{DWORD}, \c{QWORD}, \c{TWORD}, \c{OWORD} or \c{YWORD}), but will
1841 give them the smallest possible size. The keyword \c{STRICT} can be
1842 used to inhibit optimization and force a particular operand to be
1843 emitted in the specified size. For example, with the optimizer on, and
1844 in \c{BITS 16} mode,
1848 is encoded in three bytes \c{66 6A 21}, whereas
1850 \c push strict dword 33
1852 is encoded in six bytes, with a full dword immediate operand \c{66 68
1855 With the optimizer off, the same code (six bytes) is generated whether
1856 the \c{STRICT} keyword was used or not.
1859 \H{crit} \i{Critical Expressions}
1861 Although NASM has an optional multi-pass optimizer, there are some
1862 expressions which must be resolvable on the first pass. These are
1863 called \e{Critical Expressions}.
1865 The first pass is used to determine the size of all the assembled
1866 code and data, so that the second pass, when generating all the
1867 code, knows all the symbol addresses the code refers to. So one
1868 thing NASM can't handle is code whose size depends on the value of a
1869 symbol declared after the code in question. For example,
1871 \c times (label-$) db 0
1872 \c label: db 'Where am I?'
1874 The argument to \i\c{TIMES} in this case could equally legally
1875 evaluate to anything at all; NASM will reject this example because
1876 it cannot tell the size of the \c{TIMES} line when it first sees it.
1877 It will just as firmly reject the slightly \I{paradox}paradoxical
1880 \c times (label-$+1) db 0
1881 \c label: db 'NOW where am I?'
1883 in which \e{any} value for the \c{TIMES} argument is by definition
1886 NASM rejects these examples by means of a concept called a
1887 \e{critical expression}, which is defined to be an expression whose
1888 value is required to be computable in the first pass, and which must
1889 therefore depend only on symbols defined before it. The argument to
1890 the \c{TIMES} prefix is a critical expression.
1892 \H{locallab} \i{Local Labels}
1894 NASM gives special treatment to symbols beginning with a \i{period}.
1895 A label beginning with a single period is treated as a \e{local}
1896 label, which means that it is associated with the previous non-local
1897 label. So, for example:
1899 \c label1 ; some code
1907 \c label2 ; some code
1915 In the above code fragment, each \c{JNE} instruction jumps to the
1916 line immediately before it, because the two definitions of \c{.loop}
1917 are kept separate by virtue of each being associated with the
1918 previous non-local label.
1920 This form of local label handling is borrowed from the old Amiga
1921 assembler \i{DevPac}; however, NASM goes one step further, in
1922 allowing access to local labels from other parts of the code. This
1923 is achieved by means of \e{defining} a local label in terms of the
1924 previous non-local label: the first definition of \c{.loop} above is
1925 really defining a symbol called \c{label1.loop}, and the second
1926 defines a symbol called \c{label2.loop}. So, if you really needed
1929 \c label3 ; some more code
1934 Sometimes it is useful - in a macro, for instance - to be able to
1935 define a label which can be referenced from anywhere but which
1936 doesn't interfere with the normal local-label mechanism. Such a
1937 label can't be non-local because it would interfere with subsequent
1938 definitions of, and references to, local labels; and it can't be
1939 local because the macro that defined it wouldn't know the label's
1940 full name. NASM therefore introduces a third type of label, which is
1941 probably only useful in macro definitions: if a label begins with
1942 the \I{label prefix}special prefix \i\c{..@}, then it does nothing
1943 to the local label mechanism. So you could code
1945 \c label1: ; a non-local label
1946 \c .local: ; this is really label1.local
1947 \c ..@foo: ; this is a special symbol
1948 \c label2: ; another non-local label
1949 \c .local: ; this is really label2.local
1951 \c jmp ..@foo ; this will jump three lines up
1953 NASM has the capacity to define other special symbols beginning with
1954 a double period: for example, \c{..start} is used to specify the
1955 entry point in the \c{obj} output format (see \k{dotdotstart}),
1956 \c{..imagebase} is used to find out the offset from a base address
1957 of the current image in the \c{win64} output format (see \k{win64pic}).
1958 So just keep in mind that symbols beginning with a double period are
1962 \C{preproc} The NASM \i{Preprocessor}
1964 NASM contains a powerful \i{macro processor}, which supports
1965 conditional assembly, multi-level file inclusion, two forms of macro
1966 (single-line and multi-line), and a `context stack' mechanism for
1967 extra macro power. Preprocessor directives all begin with a \c{%}
1970 The preprocessor collapses all lines which end with a backslash (\\)
1971 character into a single line. Thus:
1973 \c %define THIS_VERY_LONG_MACRO_NAME_IS_DEFINED_TO \\
1976 will work like a single-line macro without the backslash-newline
1979 \H{slmacro} \i{Single-Line Macros}
1981 \S{define} The Normal Way: \I\c{%idefine}\i\c{%define}
1983 Single-line macros are defined using the \c{%define} preprocessor
1984 directive. The definitions work in a similar way to C; so you can do
1987 \c %define ctrl 0x1F &
1988 \c %define param(a,b) ((a)+(a)*(b))
1990 \c mov byte [param(2,ebx)], ctrl 'D'
1992 which will expand to
1994 \c mov byte [(2)+(2)*(ebx)], 0x1F & 'D'
1996 When the expansion of a single-line macro contains tokens which
1997 invoke another macro, the expansion is performed at invocation time,
1998 not at definition time. Thus the code
2000 \c %define a(x) 1+b(x)
2005 will evaluate in the expected way to \c{mov ax,1+2*8}, even though
2006 the macro \c{b} wasn't defined at the time of definition of \c{a}.
2008 Macros defined with \c{%define} are \i{case sensitive}: after
2009 \c{%define foo bar}, only \c{foo} will expand to \c{bar}: \c{Foo} or
2010 \c{FOO} will not. By using \c{%idefine} instead of \c{%define} (the
2011 `i' stands for `insensitive') you can define all the case variants
2012 of a macro at once, so that \c{%idefine foo bar} would cause
2013 \c{foo}, \c{Foo}, \c{FOO}, \c{fOO} and so on all to expand to
2016 There is a mechanism which detects when a macro call has occurred as
2017 a result of a previous expansion of the same macro, to guard against
2018 \i{circular references} and infinite loops. If this happens, the
2019 preprocessor will only expand the first occurrence of the macro.
2022 \c %define a(x) 1+a(x)
2026 the macro \c{a(3)} will expand once, becoming \c{1+a(3)}, and will
2027 then expand no further. This behaviour can be useful: see \k{32c}
2028 for an example of its use.
2030 You can \I{overloading, single-line macros}overload single-line
2031 macros: if you write
2033 \c %define foo(x) 1+x
2034 \c %define foo(x,y) 1+x*y
2036 the preprocessor will be able to handle both types of macro call,
2037 by counting the parameters you pass; so \c{foo(3)} will become
2038 \c{1+3} whereas \c{foo(ebx,2)} will become \c{1+ebx*2}. However, if
2043 then no other definition of \c{foo} will be accepted: a macro with
2044 no parameters prohibits the definition of the same name as a macro
2045 \e{with} parameters, and vice versa.
2047 This doesn't prevent single-line macros being \e{redefined}: you can
2048 perfectly well define a macro with
2052 and then re-define it later in the same source file with
2056 Then everywhere the macro \c{foo} is invoked, it will be expanded
2057 according to the most recent definition. This is particularly useful
2058 when defining single-line macros with \c{%assign} (see \k{assign}).
2060 You can \i{pre-define} single-line macros using the `-d' option on
2061 the NASM command line: see \k{opt-d}.
2064 \S{xdefine} Resolving \c{%define}: \I\c{%ixdefine}\i\c{%xdefine}
2066 To have a reference to an embedded single-line macro resolved at the
2067 time that the embedding macro is \e{defined}, as opposed to when the
2068 embedding macro is \e{expanded}, you need a different mechanism to the
2069 one offered by \c{%define}. The solution is to use \c{%xdefine}, or
2070 it's \I{case sensitive}case-insensitive counterpart \c{%ixdefine}.
2072 Suppose you have the following code:
2075 \c %define isFalse isTrue
2084 In this case, \c{val1} is equal to 0, and \c{val2} is equal to 1.
2085 This is because, when a single-line macro is defined using
2086 \c{%define}, it is expanded only when it is called. As \c{isFalse}
2087 expands to \c{isTrue}, the expansion will be the current value of
2088 \c{isTrue}. The first time it is called that is 0, and the second
2091 If you wanted \c{isFalse} to expand to the value assigned to the
2092 embedded macro \c{isTrue} at the time that \c{isFalse} was defined,
2093 you need to change the above code to use \c{%xdefine}.
2095 \c %xdefine isTrue 1
2096 \c %xdefine isFalse isTrue
2097 \c %xdefine isTrue 0
2101 \c %xdefine isTrue 1
2105 Now, each time that \c{isFalse} is called, it expands to 1,
2106 as that is what the embedded macro \c{isTrue} expanded to at
2107 the time that \c{isFalse} was defined.
2110 \S{indmacro} \i{Macro Indirection}: \I\c{%[}\c{%[...]}
2112 The \c{%[...]} construct can be used to expand macros in contexts
2113 where macro expansion would otherwise not occur, including in the
2114 names other macros. For example, if you have a set of macros named
2115 \c{Foo16}, \c{Foo32} and \c{Foo64}, you could write:
2117 \c mov ax,Foo%[__BITS__] ; The Foo value
2119 to use the builtin macro \c{__BITS__} (see \k{bitsm}) to automatically
2120 select between them. Similarly, the two statements:
2122 \c %xdefine Bar Quux ; Expands due to %xdefine
2123 \c %define Bar %[Quux] ; Expands due to %[...]
2125 have, in fact, exactly the same effect.
2127 \c{%[...]} concatenates to adjacent tokens in the same way that
2128 multi-line macro parameters do, see \k{concat} for details.
2131 \S{concat%+} Concatenating Single Line Macro Tokens: \i\c{%+}
2133 Individual tokens in single line macros can be concatenated, to produce
2134 longer tokens for later processing. This can be useful if there are
2135 several similar macros that perform similar functions.
2137 Please note that a space is required after \c{%+}, in order to
2138 disambiguate it from the syntax \c{%+1} used in multiline macros.
2140 As an example, consider the following:
2142 \c %define BDASTART 400h ; Start of BIOS data area
2144 \c struc tBIOSDA ; its structure
2150 Now, if we need to access the elements of tBIOSDA in different places,
2153 \c mov ax,BDASTART + tBIOSDA.COM1addr
2154 \c mov bx,BDASTART + tBIOSDA.COM2addr
2156 This will become pretty ugly (and tedious) if used in many places, and
2157 can be reduced in size significantly by using the following macro:
2159 \c ; Macro to access BIOS variables by their names (from tBDA):
2161 \c %define BDA(x) BDASTART + tBIOSDA. %+ x
2163 Now the above code can be written as:
2165 \c mov ax,BDA(COM1addr)
2166 \c mov bx,BDA(COM2addr)
2168 Using this feature, we can simplify references to a lot of macros (and,
2169 in turn, reduce typing errors).
2172 \S{selfref%?} The Macro Name Itself: \i\c{%?} and \i\c{%??}
2174 The special symbols \c{%?} and \c{%??} can be used to reference the
2175 macro name itself inside a macro expansion, this is supported for both
2176 single-and multi-line macros. \c{%?} refers to the macro name as
2177 \e{invoked}, whereas \c{%??} refers to the macro name as
2178 \e{declared}. The two are always the same for case-sensitive
2179 macros, but for case-insensitive macros, they can differ.
2183 \c %idefine Foo mov %?,%??
2195 \c %idefine keyword $%?
2197 can be used to make a keyword "disappear", for example in case a new
2198 instruction has been used as a label in older code. For example:
2200 \c %idefine pause $%? ; Hide the PAUSE instruction
2203 \S{undef} Undefining Single-Line Macros: \i\c{%undef}
2205 Single-line macros can be removed with the \c{%undef} directive. For
2206 example, the following sequence:
2213 will expand to the instruction \c{mov eax, foo}, since after
2214 \c{%undef} the macro \c{foo} is no longer defined.
2216 Macros that would otherwise be pre-defined can be undefined on the
2217 command-line using the `-u' option on the NASM command line: see
2221 \S{assign} \i{Preprocessor Variables}: \i\c{%assign}
2223 An alternative way to define single-line macros is by means of the
2224 \c{%assign} command (and its \I{case sensitive}case-insensitive
2225 counterpart \i\c{%iassign}, which differs from \c{%assign} in
2226 exactly the same way that \c{%idefine} differs from \c{%define}).
2228 \c{%assign} is used to define single-line macros which take no
2229 parameters and have a numeric value. This value can be specified in
2230 the form of an expression, and it will be evaluated once, when the
2231 \c{%assign} directive is processed.
2233 Like \c{%define}, macros defined using \c{%assign} can be re-defined
2234 later, so you can do things like
2238 to increment the numeric value of a macro.
2240 \c{%assign} is useful for controlling the termination of \c{%rep}
2241 preprocessor loops: see \k{rep} for an example of this. Another
2242 use for \c{%assign} is given in \k{16c} and \k{32c}.
2244 The expression passed to \c{%assign} is a \i{critical expression}
2245 (see \k{crit}), and must also evaluate to a pure number (rather than
2246 a relocatable reference such as a code or data address, or anything
2247 involving a register).
2250 \S{defstr} Defining Strings: \I\c{%idefstr}\i\c{%defstr}
2252 \c{%defstr}, and its case-insensitive counterpart \c{%idefstr}, define
2253 or redefine a single-line macro without parameters but converts the
2254 entire right-hand side, after macro expansion, to a quoted string
2259 \c %defstr test TEST
2263 \c %define test 'TEST'
2265 This can be used, for example, with the \c{%!} construct (see
2268 \c %defstr PATH %!PATH ; The operating system PATH variable
2271 \S{deftok} Defining Tokens: \I\c{%ideftok}\i\c{%deftok}
2273 \c{%deftok}, and its case-insensitive counterpart \c{%ideftok}, define
2274 or redefine a single-line macro without parameters but converts the
2275 second parameter, after string conversion, to a sequence of tokens.
2279 \c %deftok test 'TEST'
2283 \c %define test TEST
2286 \H{strlen} \i{String Manipulation in Macros}
2288 It's often useful to be able to handle strings in macros. NASM
2289 supports a few simple string handling macro operators from which
2290 more complex operations can be constructed.
2292 All the string operators define or redefine a value (either a string
2293 or a numeric value) to a single-line macro. When producing a string
2294 value, it may change the style of quoting of the input string or
2295 strings, and possibly use \c{\\}-escapes inside \c{`}-quoted strings.
2297 \S{strcat} \i{Concatenating Strings}: \i\c{%strcat}
2299 The \c{%strcat} operator concatenates quoted strings and assign them to
2300 a single-line macro.
2304 \c %strcat alpha "Alpha: ", '12" screen'
2306 ... would assign the value \c{'Alpha: 12" screen'} to \c{alpha}.
2309 \c %strcat beta '"foo"\', "'bar'"
2311 ... would assign the value \c{`"foo"\\\\'bar'`} to \c{beta}.
2313 The use of commas to separate strings is permitted but optional.
2316 \S{strlen} \i{String Length}: \i\c{%strlen}
2318 The \c{%strlen} operator assigns the length of a string to a macro.
2321 \c %strlen charcnt 'my string'
2323 In this example, \c{charcnt} would receive the value 9, just as
2324 if an \c{%assign} had been used. In this example, \c{'my string'}
2325 was a literal string but it could also have been a single-line
2326 macro that expands to a string, as in the following example:
2328 \c %define sometext 'my string'
2329 \c %strlen charcnt sometext
2331 As in the first case, this would result in \c{charcnt} being
2332 assigned the value of 9.
2335 \S{substr} \i{Extracting Substrings}: \i\c{%substr}
2337 Individual letters or substrings in strings can be extracted using the
2338 \c{%substr} operator. An example of its use is probably more useful
2339 than the description:
2341 \c %substr mychar 'xyzw' 1 ; equivalent to %define mychar 'x'
2342 \c %substr mychar 'xyzw' 2 ; equivalent to %define mychar 'y'
2343 \c %substr mychar 'xyzw' 3 ; equivalent to %define mychar 'z'
2344 \c %substr mychar 'xyzw' 2,2 ; equivalent to %define mychar 'yz'
2345 \c %substr mychar 'xyzw' 2,-1 ; equivalent to %define mychar 'yzw'
2346 \c %substr mychar 'xyzw' 2,-2 ; equivalent to %define mychar 'yz'
2348 As with \c{%strlen} (see \k{strlen}), the first parameter is the
2349 single-line macro to be created and the second is the string. The
2350 third parameter specifies the first character to be selected, and the
2351 optional fourth parameter preceeded by comma) is the length. Note
2352 that the first index is 1, not 0 and the last index is equal to the
2353 value that \c{%strlen} would assign given the same string. Index
2354 values out of range result in an empty string. A negative length
2355 means "until N-1 characters before the end of string", i.e. \c{-1}
2356 means until end of string, \c{-2} until one character before, etc.
2359 \H{mlmacro} \i{Multi-Line Macros}: \I\c{%imacro}\i\c{%macro}
2361 Multi-line macros are much more like the type of macro seen in MASM
2362 and TASM: a multi-line macro definition in NASM looks something like
2365 \c %macro prologue 1
2373 This defines a C-like function prologue as a macro: so you would
2374 invoke the macro with a call such as
2376 \c myfunc: prologue 12
2378 which would expand to the three lines of code
2384 The number \c{1} after the macro name in the \c{%macro} line defines
2385 the number of parameters the macro \c{prologue} expects to receive.
2386 The use of \c{%1} inside the macro definition refers to the first
2387 parameter to the macro call. With a macro taking more than one
2388 parameter, subsequent parameters would be referred to as \c{%2},
2391 Multi-line macros, like single-line macros, are \i{case-sensitive},
2392 unless you define them using the alternative directive \c{%imacro}.
2394 If you need to pass a comma as \e{part} of a parameter to a
2395 multi-line macro, you can do that by enclosing the entire parameter
2396 in \I{braces, around macro parameters}braces. So you could code
2405 \c silly 'a', letter_a ; letter_a: db 'a'
2406 \c silly 'ab', string_ab ; string_ab: db 'ab'
2407 \c silly {13,10}, crlf ; crlf: db 13,10
2410 \#\S{mlrmacro} \i{Recursive Multi-Line Macros}: \I\c{%irmacro}\i\c{%rmacro}
2412 \#A multi-line macro cannot be referenced within itself, in order to
2413 \#prevent accidental infinite recursion.
2415 \#Recursive multi-line macros allow for self-referencing, with the
2416 \#caveat that the user is aware of the existence, use and purpose of
2417 \#recursive multi-line macros. There is also a generous, but sane, upper
2418 \#limit to the number of recursions, in order to prevent run-away memory
2419 \#consumption in case of accidental infinite recursion.
2421 \#As with non-recursive multi-line macros, recursive multi-line macros are
2422 \#\i{case-sensitive}, unless you define them using the alternative
2423 \#directive \c{%irmacro}.
2425 \S{mlmacover} Overloading Multi-Line Macros\I{overloading, multi-line macros}
2427 As with single-line macros, multi-line macros can be overloaded by
2428 defining the same macro name several times with different numbers of
2429 parameters. This time, no exception is made for macros with no
2430 parameters at all. So you could define
2432 \c %macro prologue 0
2439 to define an alternative form of the function prologue which
2440 allocates no local stack space.
2442 Sometimes, however, you might want to `overload' a machine
2443 instruction; for example, you might want to define
2452 so that you could code
2454 \c push ebx ; this line is not a macro call
2455 \c push eax,ecx ; but this one is
2457 Ordinarily, NASM will give a warning for the first of the above two
2458 lines, since \c{push} is now defined to be a macro, and is being
2459 invoked with a number of parameters for which no definition has been
2460 given. The correct code will still be generated, but the assembler
2461 will give a warning. This warning can be disabled by the use of the
2462 \c{-w-macro-params} command-line option (see \k{opt-w}).
2465 \S{maclocal} \i{Macro-Local Labels}
2467 NASM allows you to define labels within a multi-line macro
2468 definition in such a way as to make them local to the macro call: so
2469 calling the same macro multiple times will use a different label
2470 each time. You do this by prefixing \i\c{%%} to the label name. So
2471 you can invent an instruction which executes a \c{RET} if the \c{Z}
2472 flag is set by doing this:
2482 You can call this macro as many times as you want, and every time
2483 you call it NASM will make up a different `real' name to substitute
2484 for the label \c{%%skip}. The names NASM invents are of the form
2485 \c{..@2345.skip}, where the number 2345 changes with every macro
2486 call. The \i\c{..@} prefix prevents macro-local labels from
2487 interfering with the local label mechanism, as described in
2488 \k{locallab}. You should avoid defining your own labels in this form
2489 (the \c{..@} prefix, then a number, then another period) in case
2490 they interfere with macro-local labels.
2493 \S{mlmacgre} \i{Greedy Macro Parameters}
2495 Occasionally it is useful to define a macro which lumps its entire
2496 command line into one parameter definition, possibly after
2497 extracting one or two smaller parameters from the front. An example
2498 might be a macro to write a text string to a file in MS-DOS, where
2499 you might want to be able to write
2501 \c writefile [filehandle],"hello, world",13,10
2503 NASM allows you to define the last parameter of a macro to be
2504 \e{greedy}, meaning that if you invoke the macro with more
2505 parameters than it expects, all the spare parameters get lumped into
2506 the last defined one along with the separating commas. So if you
2509 \c %macro writefile 2+
2515 \c mov cx,%%endstr-%%str
2522 then the example call to \c{writefile} above will work as expected:
2523 the text before the first comma, \c{[filehandle]}, is used as the
2524 first macro parameter and expanded when \c{%1} is referred to, and
2525 all the subsequent text is lumped into \c{%2} and placed after the
2528 The greedy nature of the macro is indicated to NASM by the use of
2529 the \I{+ modifier}\c{+} sign after the parameter count on the
2532 If you define a greedy macro, you are effectively telling NASM how
2533 it should expand the macro given \e{any} number of parameters from
2534 the actual number specified up to infinity; in this case, for
2535 example, NASM now knows what to do when it sees a call to
2536 \c{writefile} with 2, 3, 4 or more parameters. NASM will take this
2537 into account when overloading macros, and will not allow you to
2538 define another form of \c{writefile} taking 4 parameters (for
2541 Of course, the above macro could have been implemented as a
2542 non-greedy macro, in which case the call to it would have had to
2545 \c writefile [filehandle], {"hello, world",13,10}
2547 NASM provides both mechanisms for putting \i{commas in macro
2548 parameters}, and you choose which one you prefer for each macro
2551 See \k{sectmac} for a better way to write the above macro.
2554 \S{mlmacdef} \i{Default Macro Parameters}
2556 NASM also allows you to define a multi-line macro with a \e{range}
2557 of allowable parameter counts. If you do this, you can specify
2558 defaults for \i{omitted parameters}. So, for example:
2560 \c %macro die 0-1 "Painful program death has occurred."
2568 This macro (which makes use of the \c{writefile} macro defined in
2569 \k{mlmacgre}) can be called with an explicit error message, which it
2570 will display on the error output stream before exiting, or it can be
2571 called with no parameters, in which case it will use the default
2572 error message supplied in the macro definition.
2574 In general, you supply a minimum and maximum number of parameters
2575 for a macro of this type; the minimum number of parameters are then
2576 required in the macro call, and then you provide defaults for the
2577 optional ones. So if a macro definition began with the line
2579 \c %macro foobar 1-3 eax,[ebx+2]
2581 then it could be called with between one and three parameters, and
2582 \c{%1} would always be taken from the macro call. \c{%2}, if not
2583 specified by the macro call, would default to \c{eax}, and \c{%3} if
2584 not specified would default to \c{[ebx+2]}.
2586 You can provide extra information to a macro by providing
2587 too many default parameters:
2589 \c %macro quux 1 something
2591 This will trigger a warning by default; see \k{opt-w} for
2593 When \c{quux} is invoked, it receives not one but two parameters.
2594 \c{something} can be referred to as \c{%2}. The difference
2595 between passing \c{something} this way and writing \c{something}
2596 in the macro body is that with this way \c{something} is evaluated
2597 when the macro is defined, not when it is expanded.
2599 You may omit parameter defaults from the macro definition, in which
2600 case the parameter default is taken to be blank. This can be useful
2601 for macros which can take a variable number of parameters, since the
2602 \i\c{%0} token (see \k{percent0}) allows you to determine how many
2603 parameters were really passed to the macro call.
2605 This defaulting mechanism can be combined with the greedy-parameter
2606 mechanism; so the \c{die} macro above could be made more powerful,
2607 and more useful, by changing the first line of the definition to
2609 \c %macro die 0-1+ "Painful program death has occurred.",13,10
2611 The maximum parameter count can be infinite, denoted by \c{*}. In
2612 this case, of course, it is impossible to provide a \e{full} set of
2613 default parameters. Examples of this usage are shown in \k{rotate}.
2616 \S{percent0} \i\c{%0}: \I{counting macro parameters}Macro Parameter Counter
2618 The parameter reference \c{%0} will return a numeric constant giving the
2619 number of parameters received, that is, if \c{%0} is n then \c{%}n is the
2620 last parameter. \c{%0} is mostly useful for macros that can take a variable
2621 number of parameters. It can be used as an argument to \c{%rep}
2622 (see \k{rep}) in order to iterate through all the parameters of a macro.
2623 Examples are given in \k{rotate}.
2626 \S{rotate} \i\c{%rotate}: \i{Rotating Macro Parameters}
2628 Unix shell programmers will be familiar with the \I{shift
2629 command}\c{shift} shell command, which allows the arguments passed
2630 to a shell script (referenced as \c{$1}, \c{$2} and so on) to be
2631 moved left by one place, so that the argument previously referenced
2632 as \c{$2} becomes available as \c{$1}, and the argument previously
2633 referenced as \c{$1} is no longer available at all.
2635 NASM provides a similar mechanism, in the form of \c{%rotate}. As
2636 its name suggests, it differs from the Unix \c{shift} in that no
2637 parameters are lost: parameters rotated off the left end of the
2638 argument list reappear on the right, and vice versa.
2640 \c{%rotate} is invoked with a single numeric argument (which may be
2641 an expression). The macro parameters are rotated to the left by that
2642 many places. If the argument to \c{%rotate} is negative, the macro
2643 parameters are rotated to the right.
2645 \I{iterating over macro parameters}So a pair of macros to save and
2646 restore a set of registers might work as follows:
2648 \c %macro multipush 1-*
2657 This macro invokes the \c{PUSH} instruction on each of its arguments
2658 in turn, from left to right. It begins by pushing its first
2659 argument, \c{%1}, then invokes \c{%rotate} to move all the arguments
2660 one place to the left, so that the original second argument is now
2661 available as \c{%1}. Repeating this procedure as many times as there
2662 were arguments (achieved by supplying \c{%0} as the argument to
2663 \c{%rep}) causes each argument in turn to be pushed.
2665 Note also the use of \c{*} as the maximum parameter count,
2666 indicating that there is no upper limit on the number of parameters
2667 you may supply to the \i\c{multipush} macro.
2669 It would be convenient, when using this macro, to have a \c{POP}
2670 equivalent, which \e{didn't} require the arguments to be given in
2671 reverse order. Ideally, you would write the \c{multipush} macro
2672 call, then cut-and-paste the line to where the pop needed to be
2673 done, and change the name of the called macro to \c{multipop}, and
2674 the macro would take care of popping the registers in the opposite
2675 order from the one in which they were pushed.
2677 This can be done by the following definition:
2679 \c %macro multipop 1-*
2688 This macro begins by rotating its arguments one place to the
2689 \e{right}, so that the original \e{last} argument appears as \c{%1}.
2690 This is then popped, and the arguments are rotated right again, so
2691 the second-to-last argument becomes \c{%1}. Thus the arguments are
2692 iterated through in reverse order.
2695 \S{concat} \i{Concatenating Macro Parameters}
2697 NASM can concatenate macro parameters and macro indirection constructs
2698 on to other text surrounding them. This allows you to declare a family
2699 of symbols, for example, in a macro definition. If, for example, you
2700 wanted to generate a table of key codes along with offsets into the
2701 table, you could code something like
2703 \c %macro keytab_entry 2
2705 \c keypos%1 equ $-keytab
2711 \c keytab_entry F1,128+1
2712 \c keytab_entry F2,128+2
2713 \c keytab_entry Return,13
2715 which would expand to
2718 \c keyposF1 equ $-keytab
2720 \c keyposF2 equ $-keytab
2722 \c keyposReturn equ $-keytab
2725 You can just as easily concatenate text on to the other end of a
2726 macro parameter, by writing \c{%1foo}.
2728 If you need to append a \e{digit} to a macro parameter, for example
2729 defining labels \c{foo1} and \c{foo2} when passed the parameter
2730 \c{foo}, you can't code \c{%11} because that would be taken as the
2731 eleventh macro parameter. Instead, you must code
2732 \I{braces, after % sign}\c{%\{1\}1}, which will separate the first
2733 \c{1} (giving the number of the macro parameter) from the second
2734 (literal text to be concatenated to the parameter).
2736 This concatenation can also be applied to other preprocessor in-line
2737 objects, such as macro-local labels (\k{maclocal}) and context-local
2738 labels (\k{ctxlocal}). In all cases, ambiguities in syntax can be
2739 resolved by enclosing everything after the \c{%} sign and before the
2740 literal text in braces: so \c{%\{%foo\}bar} concatenates the text
2741 \c{bar} to the end of the real name of the macro-local label
2742 \c{%%foo}. (This is unnecessary, since the form NASM uses for the
2743 real names of macro-local labels means that the two usages
2744 \c{%\{%foo\}bar} and \c{%%foobar} would both expand to the same
2745 thing anyway; nevertheless, the capability is there.)
2747 The single-line macro indirection construct, \c{%[...]}
2748 (\k{indmacro}), behaves the same way as macro parameters for the
2749 purpose of concatenation.
2751 See also the \c{%+} operator, \k{concat%+}.
2754 \S{mlmaccc} \i{Condition Codes as Macro Parameters}
2756 NASM can give special treatment to a macro parameter which contains
2757 a condition code. For a start, you can refer to the macro parameter
2758 \c{%1} by means of the alternative syntax \i\c{%+1}, which informs
2759 NASM that this macro parameter is supposed to contain a condition
2760 code, and will cause the preprocessor to report an error message if
2761 the macro is called with a parameter which is \e{not} a valid
2764 Far more usefully, though, you can refer to the macro parameter by
2765 means of \i\c{%-1}, which NASM will expand as the \e{inverse}
2766 condition code. So the \c{retz} macro defined in \k{maclocal} can be
2767 replaced by a general \i{conditional-return macro} like this:
2777 This macro can now be invoked using calls like \c{retc ne}, which
2778 will cause the conditional-jump instruction in the macro expansion
2779 to come out as \c{JE}, or \c{retc po} which will make the jump a
2782 The \c{%+1} macro-parameter reference is quite happy to interpret
2783 the arguments \c{CXZ} and \c{ECXZ} as valid condition codes;
2784 however, \c{%-1} will report an error if passed either of these,
2785 because no inverse condition code exists.
2788 \S{nolist} \i{Disabling Listing Expansion}\I\c{.nolist}
2790 When NASM is generating a listing file from your program, it will
2791 generally expand multi-line macros by means of writing the macro
2792 call and then listing each line of the expansion. This allows you to
2793 see which instructions in the macro expansion are generating what
2794 code; however, for some macros this clutters the listing up
2797 NASM therefore provides the \c{.nolist} qualifier, which you can
2798 include in a macro definition to inhibit the expansion of the macro
2799 in the listing file. The \c{.nolist} qualifier comes directly after
2800 the number of parameters, like this:
2802 \c %macro foo 1.nolist
2806 \c %macro bar 1-5+.nolist a,b,c,d,e,f,g,h
2808 \S{unmacro} Undefining Multi-Line Macros: \i\c{%unmacro}
2810 Multi-line macros can be removed with the \c{%unmacro} directive.
2811 Unlike the \c{%undef} directive, however, \c{%unmacro} takes an
2812 argument specification, and will only remove \i{exact matches} with
2813 that argument specification.
2822 removes the previously defined macro \c{foo}, but
2829 does \e{not} remove the macro \c{bar}, since the argument
2830 specification does not match exactly.
2833 \#\S{exitmacro} Exiting Multi-Line Macros: \i\c{%exitmacro}
2835 \#Multi-line macro expansions can be arbitrarily terminated with
2836 \#the \c{%exitmacro} directive.
2848 \H{condasm} \i{Conditional Assembly}\I\c{%if}
2850 Similarly to the C preprocessor, NASM allows sections of a source
2851 file to be assembled only if certain conditions are met. The general
2852 syntax of this feature looks like this:
2855 \c ; some code which only appears if <condition> is met
2856 \c %elif<condition2>
2857 \c ; only appears if <condition> is not met but <condition2> is
2859 \c ; this appears if neither <condition> nor <condition2> was met
2862 The inverse forms \i\c{%ifn} and \i\c{%elifn} are also supported.
2864 The \i\c{%else} clause is optional, as is the \i\c{%elif} clause.
2865 You can have more than one \c{%elif} clause as well.
2867 There are a number of variants of the \c{%if} directive. Each has its
2868 corresponding \c{%elif}, \c{%ifn}, and \c{%elifn} directives; for
2869 example, the equivalents to the \c{%ifdef} directive are \c{%elifdef},
2870 \c{%ifndef}, and \c{%elifndef}.
2872 \S{ifdef} \i\c{%ifdef}: Testing Single-Line Macro Existence\I{testing,
2873 single-line macro existence}
2875 Beginning a conditional-assembly block with the line \c{%ifdef
2876 MACRO} will assemble the subsequent code if, and only if, a
2877 single-line macro called \c{MACRO} is defined. If not, then the
2878 \c{%elif} and \c{%else} blocks (if any) will be processed instead.
2880 For example, when debugging a program, you might want to write code
2883 \c ; perform some function
2885 \c writefile 2,"Function performed successfully",13,10
2887 \c ; go and do something else
2889 Then you could use the command-line option \c{-dDEBUG} to create a
2890 version of the program which produced debugging messages, and remove
2891 the option to generate the final release version of the program.
2893 You can test for a macro \e{not} being defined by using
2894 \i\c{%ifndef} instead of \c{%ifdef}. You can also test for macro
2895 definitions in \c{%elif} blocks by using \i\c{%elifdef} and
2899 \S{ifmacro} \i\c{%ifmacro}: Testing Multi-Line Macro
2900 Existence\I{testing, multi-line macro existence}
2902 The \c{%ifmacro} directive operates in the same way as the \c{%ifdef}
2903 directive, except that it checks for the existence of a multi-line macro.
2905 For example, you may be working with a large project and not have control
2906 over the macros in a library. You may want to create a macro with one
2907 name if it doesn't already exist, and another name if one with that name
2910 The \c{%ifmacro} is considered true if defining a macro with the given name
2911 and number of arguments would cause a definitions conflict. For example:
2913 \c %ifmacro MyMacro 1-3
2915 \c %error "MyMacro 1-3" causes a conflict with an existing macro.
2919 \c %macro MyMacro 1-3
2921 \c ; insert code to define the macro
2927 This will create the macro "MyMacro 1-3" if no macro already exists which
2928 would conflict with it, and emits a warning if there would be a definition
2931 You can test for the macro not existing by using the \i\c{%ifnmacro} instead
2932 of \c{%ifmacro}. Additional tests can be performed in \c{%elif} blocks by using
2933 \i\c{%elifmacro} and \i\c{%elifnmacro}.
2936 \S{ifctx} \i\c{%ifctx}: Testing the Context Stack\I{testing, context
2939 The conditional-assembly construct \c{%ifctx} will cause the
2940 subsequent code to be assembled if and only if the top context on
2941 the preprocessor's context stack has the same name as one of the arguments.
2942 As with \c{%ifdef}, the inverse and \c{%elif} forms \i\c{%ifnctx},
2943 \i\c{%elifctx} and \i\c{%elifnctx} are also supported.
2945 For more details of the context stack, see \k{ctxstack}. For a
2946 sample use of \c{%ifctx}, see \k{blockif}.
2949 \S{if} \i\c{%if}: Testing Arbitrary Numeric Expressions\I{testing,
2950 arbitrary numeric expressions}
2952 The conditional-assembly construct \c{%if expr} will cause the
2953 subsequent code to be assembled if and only if the value of the
2954 numeric expression \c{expr} is non-zero. An example of the use of
2955 this feature is in deciding when to break out of a \c{%rep}
2956 preprocessor loop: see \k{rep} for a detailed example.
2958 The expression given to \c{%if}, and its counterpart \i\c{%elif}, is
2959 a critical expression (see \k{crit}).
2961 \c{%if} extends the normal NASM expression syntax, by providing a
2962 set of \i{relational operators} which are not normally available in
2963 expressions. The operators \i\c{=}, \i\c{<}, \i\c{>}, \i\c{<=},
2964 \i\c{>=} and \i\c{<>} test equality, less-than, greater-than,
2965 less-or-equal, greater-or-equal and not-equal respectively. The
2966 C-like forms \i\c{==} and \i\c{!=} are supported as alternative
2967 forms of \c{=} and \c{<>}. In addition, low-priority logical
2968 operators \i\c{&&}, \i\c{^^} and \i\c{||} are provided, supplying
2969 \i{logical AND}, \i{logical XOR} and \i{logical OR}. These work like
2970 the C logical operators (although C has no logical XOR), in that
2971 they always return either 0 or 1, and treat any non-zero input as 1
2972 (so that \c{^^}, for example, returns 1 if exactly one of its inputs
2973 is zero, and 0 otherwise). The relational operators also return 1
2974 for true and 0 for false.
2976 Like other \c{%if} constructs, \c{%if} has a counterpart
2977 \i\c{%elif}, and negative forms \i\c{%ifn} and \i\c{%elifn}.
2979 \S{ifidn} \i\c{%ifidn} and \i\c{%ifidni}: Testing Exact Text
2980 Identity\I{testing, exact text identity}
2982 The construct \c{%ifidn text1,text2} will cause the subsequent code
2983 to be assembled if and only if \c{text1} and \c{text2}, after
2984 expanding single-line macros, are identical pieces of text.
2985 Differences in white space are not counted.
2987 \c{%ifidni} is similar to \c{%ifidn}, but is \i{case-insensitive}.
2989 For example, the following macro pushes a register or number on the
2990 stack, and allows you to treat \c{IP} as a real register:
2992 \c %macro pushparam 1
3003 Like other \c{%if} constructs, \c{%ifidn} has a counterpart
3004 \i\c{%elifidn}, and negative forms \i\c{%ifnidn} and \i\c{%elifnidn}.
3005 Similarly, \c{%ifidni} has counterparts \i\c{%elifidni},
3006 \i\c{%ifnidni} and \i\c{%elifnidni}.
3008 \S{iftyp} \i\c{%ifid}, \i\c{%ifnum}, \i\c{%ifstr}: Testing Token
3009 Types\I{testing, token types}
3011 Some macros will want to perform different tasks depending on
3012 whether they are passed a number, a string, or an identifier. For
3013 example, a string output macro might want to be able to cope with
3014 being passed either a string constant or a pointer to an existing
3017 The conditional assembly construct \c{%ifid}, taking one parameter
3018 (which may be blank), assembles the subsequent code if and only if
3019 the first token in the parameter exists and is an identifier.
3020 \c{%ifnum} works similarly, but tests for the token being a numeric
3021 constant; \c{%ifstr} tests for it being a string.
3023 For example, the \c{writefile} macro defined in \k{mlmacgre} can be
3024 extended to take advantage of \c{%ifstr} in the following fashion:
3026 \c %macro writefile 2-3+
3035 \c %%endstr: mov dx,%%str
3036 \c mov cx,%%endstr-%%str
3047 Then the \c{writefile} macro can cope with being called in either of
3048 the following two ways:
3050 \c writefile [file], strpointer, length
3051 \c writefile [file], "hello", 13, 10
3053 In the first, \c{strpointer} is used as the address of an
3054 already-declared string, and \c{length} is used as its length; in
3055 the second, a string is given to the macro, which therefore declares
3056 it itself and works out the address and length for itself.
3058 Note the use of \c{%if} inside the \c{%ifstr}: this is to detect
3059 whether the macro was passed two arguments (so the string would be a
3060 single string constant, and \c{db %2} would be adequate) or more (in
3061 which case, all but the first two would be lumped together into
3062 \c{%3}, and \c{db %2,%3} would be required).
3064 The usual \I\c{%elifid}\I\c{%elifnum}\I\c{%elifstr}\c{%elif}...,
3065 \I\c{%ifnid}\I\c{%ifnnum}\I\c{%ifnstr}\c{%ifn}..., and
3066 \I\c{%elifnid}\I\c{%elifnnum}\I\c{%elifnstr}\c{%elifn}... versions
3067 exist for each of \c{%ifid}, \c{%ifnum} and \c{%ifstr}.
3069 \S{iftoken} \i\c{%iftoken}: Test for a Single Token
3071 Some macros will want to do different things depending on if it is
3072 passed a single token (e.g. paste it to something else using \c{%+})
3073 versus a multi-token sequence.
3075 The conditional assembly construct \c{%iftoken} assembles the
3076 subsequent code if and only if the expanded parameters consist of
3077 exactly one token, possibly surrounded by whitespace.
3083 will assemble the subsequent code, but
3087 will not, since \c{-1} contains two tokens: the unary minus operator
3088 \c{-}, and the number \c{1}.
3090 The usual \i\c{%eliftoken}, \i\c\{%ifntoken}, and \i\c{%elifntoken}
3091 variants are also provided.
3093 \S{ifempty} \i\c{%ifempty}: Test for Empty Expansion
3095 The conditional assembly construct \c{%ifempty} assembles the
3096 subsequent code if and only if the expanded parameters do not contain
3097 any tokens at all, whitespace excepted.
3099 The usual \i\c{%elifempty}, \i\c\{%ifnempty}, and \i\c{%elifnempty}
3100 variants are also provided.
3102 \H{rep} \i{Preprocessor Loops}\I{repeating code}: \i\c{%rep}
3104 NASM's \c{TIMES} prefix, though useful, cannot be used to invoke a
3105 multi-line macro multiple times, because it is processed by NASM
3106 after macros have already been expanded. Therefore NASM provides
3107 another form of loop, this time at the preprocessor level: \c{%rep}.
3109 The directives \c{%rep} and \i\c{%endrep} (\c{%rep} takes a numeric
3110 argument, which can be an expression; \c{%endrep} takes no
3111 arguments) can be used to enclose a chunk of code, which is then
3112 replicated as many times as specified by the preprocessor:
3116 \c inc word [table+2*i]
3120 This will generate a sequence of 64 \c{INC} instructions,
3121 incrementing every word of memory from \c{[table]} to
3124 For more complex termination conditions, or to break out of a repeat
3125 loop part way along, you can use the \i\c{%exitrep} directive to
3126 terminate the loop, like this:
3141 \c fib_number equ ($-fibonacci)/2
3143 This produces a list of all the Fibonacci numbers that will fit in
3144 16 bits. Note that a maximum repeat count must still be given to
3145 \c{%rep}. This is to prevent the possibility of NASM getting into an
3146 infinite loop in the preprocessor, which (on multitasking or
3147 multi-user systems) would typically cause all the system memory to
3148 be gradually used up and other applications to start crashing.
3151 \H{files} Source Files and Dependencies
3153 These commands allow you to split your sources into multiple files.
3155 \S{include} \i\c{%include}: \i{Including Other Files}
3157 Using, once again, a very similar syntax to the C preprocessor,
3158 NASM's preprocessor lets you include other source files into your
3159 code. This is done by the use of the \i\c{%include} directive:
3161 \c %include "macros.mac"
3163 will include the contents of the file \c{macros.mac} into the source
3164 file containing the \c{%include} directive.
3166 Include files are \I{searching for include files}searched for in the
3167 current directory (the directory you're in when you run NASM, as
3168 opposed to the location of the NASM executable or the location of
3169 the source file), plus any directories specified on the NASM command
3170 line using the \c{-i} option.
3172 The standard C idiom for preventing a file being included more than
3173 once is just as applicable in NASM: if the file \c{macros.mac} has
3176 \c %ifndef MACROS_MAC
3177 \c %define MACROS_MAC
3178 \c ; now define some macros
3181 then including the file more than once will not cause errors,
3182 because the second time the file is included nothing will happen
3183 because the macro \c{MACROS_MAC} will already be defined.
3185 You can force a file to be included even if there is no \c{%include}
3186 directive that explicitly includes it, by using the \i\c{-p} option
3187 on the NASM command line (see \k{opt-p}).
3190 \S{pathsearch} \i\c{%pathsearch}: Search the Include Path
3192 The \c{%pathsearch} directive takes a single-line macro name and a
3193 filename, and declare or redefines the specified single-line macro to
3194 be the include-path-resolved version of the filename, if the file
3195 exists (otherwise, it is passed unchanged.)
3199 \c %pathsearch MyFoo "foo.bin"
3201 ... with \c{-Ibins/} in the include path may end up defining the macro
3202 \c{MyFoo} to be \c{"bins/foo.bin"}.
3205 \S{depend} \i\c{%depend}: Add Dependent Files
3207 The \c{%depend} directive takes a filename and adds it to the list of
3208 files to be emitted as dependency generation when the \c{-M} options
3209 and its relatives (see \k{opt-M}) are used. It produces no output.
3211 This is generally used in conjunction with \c{%pathsearch}. For
3212 example, a simplified version of the standard macro wrapper for the
3213 \c{INCBIN} directive looks like:
3215 \c %imacro incbin 1-2+ 0
3216 \c %pathsearch dep %1
3221 This first resolves the location of the file into the macro \c{dep},
3222 then adds it to the dependency lists, and finally issues the
3223 assembler-level \c{INCBIN} directive.
3226 \S{use} \i\c{%use}: Include Standard Macro Package
3228 The \c{%use} directive is similar to \c{%include}, but rather than
3229 including the contents of a file, it includes a named standard macro
3230 package. The standard macro packages are part of NASM, and are
3231 described in \k{macropkg}.
3233 Unlike the \c{%include} directive, package names for the \c{%use}
3234 directive do not require quotes, but quotes are permitted. In NASM
3235 2.04 and 2.05 the unquoted form would be macro-expanded; this is no
3236 longer true. Thus, the following lines are equivalent:
3241 Standard macro packages are protected from multiple inclusion. When a
3242 standard macro package is used, a testable single-line macro of the
3243 form \c{__USE_}\e{package}\c{__} is also defined, see \k{use_def}.
3245 \H{ctxstack} The \i{Context Stack}
3247 Having labels that are local to a macro definition is sometimes not
3248 quite powerful enough: sometimes you want to be able to share labels
3249 between several macro calls. An example might be a \c{REPEAT} ...
3250 \c{UNTIL} loop, in which the expansion of the \c{REPEAT} macro
3251 would need to be able to refer to a label which the \c{UNTIL} macro
3252 had defined. However, for such a macro you would also want to be
3253 able to nest these loops.
3255 NASM provides this level of power by means of a \e{context stack}.
3256 The preprocessor maintains a stack of \e{contexts}, each of which is
3257 characterized by a name. You add a new context to the stack using
3258 the \i\c{%push} directive, and remove one using \i\c{%pop}. You can
3259 define labels that are local to a particular context on the stack.
3262 \S{pushpop} \i\c{%push} and \i\c{%pop}: \I{creating
3263 contexts}\I{removing contexts}Creating and Removing Contexts
3265 The \c{%push} directive is used to create a new context and place it
3266 on the top of the context stack. \c{%push} takes an optional argument,
3267 which is the name of the context. For example:
3271 This pushes a new context called \c{foobar} on the stack. You can have
3272 several contexts on the stack with the same name: they can still be
3273 distinguished. If no name is given, the context is unnamed (this is
3274 normally used when both the \c{%push} and the \c{%pop} are inside a
3275 single macro definition.)
3277 The directive \c{%pop}, taking one optional argument, removes the top
3278 context from the context stack and destroys it, along with any
3279 labels associated with it. If an argument is given, it must match the
3280 name of the current context, otherwise it will issue an error.
3283 \S{ctxlocal} \i{Context-Local Labels}
3285 Just as the usage \c{%%foo} defines a label which is local to the
3286 particular macro call in which it is used, the usage \I{%$}\c{%$foo}
3287 is used to define a label which is local to the context on the top
3288 of the context stack. So the \c{REPEAT} and \c{UNTIL} example given
3289 above could be implemented by means of:
3305 and invoked by means of, for example,
3313 which would scan every fourth byte of a string in search of the byte
3316 If you need to define, or access, labels local to the context
3317 \e{below} the top one on the stack, you can use \I{%$$}\c{%$$foo}, or
3318 \c{%$$$foo} for the context below that, and so on.
3321 \S{ctxdefine} \i{Context-Local Single-Line Macros}
3323 NASM also allows you to define single-line macros which are local to
3324 a particular context, in just the same way:
3326 \c %define %$localmac 3
3328 will define the single-line macro \c{%$localmac} to be local to the
3329 top context on the stack. Of course, after a subsequent \c{%push},
3330 it can then still be accessed by the name \c{%$$localmac}.
3333 \S{ctxrepl} \i\c{%repl}: \I{renaming contexts}Renaming a Context
3335 If you need to change the name of the top context on the stack (in
3336 order, for example, to have it respond differently to \c{%ifctx}),
3337 you can execute a \c{%pop} followed by a \c{%push}; but this will
3338 have the side effect of destroying all context-local labels and
3339 macros associated with the context that was just popped.
3341 NASM provides the directive \c{%repl}, which \e{replaces} a context
3342 with a different name, without touching the associated macros and
3343 labels. So you could replace the destructive code
3348 with the non-destructive version \c{%repl newname}.
3351 \S{blockif} Example Use of the \i{Context Stack}: \i{Block IFs}
3353 This example makes use of almost all the context-stack features,
3354 including the conditional-assembly construct \i\c{%ifctx}, to
3355 implement a block IF statement as a set of macros.
3371 \c %error "expected `if' before `else'"
3385 \c %error "expected `if' or `else' before `endif'"
3390 This code is more robust than the \c{REPEAT} and \c{UNTIL} macros
3391 given in \k{ctxlocal}, because it uses conditional assembly to check
3392 that the macros are issued in the right order (for example, not
3393 calling \c{endif} before \c{if}) and issues a \c{%error} if they're
3396 In addition, the \c{endif} macro has to be able to cope with the two
3397 distinct cases of either directly following an \c{if}, or following
3398 an \c{else}. It achieves this, again, by using conditional assembly
3399 to do different things depending on whether the context on top of
3400 the stack is \c{if} or \c{else}.
3402 The \c{else} macro has to preserve the context on the stack, in
3403 order to have the \c{%$ifnot} referred to by the \c{if} macro be the
3404 same as the one defined by the \c{endif} macro, but has to change
3405 the context's name so that \c{endif} will know there was an
3406 intervening \c{else}. It does this by the use of \c{%repl}.
3408 A sample usage of these macros might look like:
3430 The block-\c{IF} macros handle nesting quite happily, by means of
3431 pushing another context, describing the inner \c{if}, on top of the
3432 one describing the outer \c{if}; thus \c{else} and \c{endif} always
3433 refer to the last unmatched \c{if} or \c{else}.
3436 \H{stackrel} \i{Stack Relative Preprocessor Directives}
3438 The following preprocessor directives provide a way to use
3439 labels to refer to local variables allocated on the stack.
3441 \b\c{%arg} (see \k{arg})
3443 \b\c{%stacksize} (see \k{stacksize})
3445 \b\c{%local} (see \k{local})
3448 \S{arg} \i\c{%arg} Directive
3450 The \c{%arg} directive is used to simplify the handling of
3451 parameters passed on the stack. Stack based parameter passing
3452 is used by many high level languages, including C, C++ and Pascal.
3454 While NASM has macros which attempt to duplicate this
3455 functionality (see \k{16cmacro}), the syntax is not particularly
3456 convenient to use and is not TASM compatible. Here is an example
3457 which shows the use of \c{%arg} without any external macros:
3461 \c %push mycontext ; save the current context
3462 \c %stacksize large ; tell NASM to use bp
3463 \c %arg i:word, j_ptr:word
3470 \c %pop ; restore original context
3472 This is similar to the procedure defined in \k{16cmacro} and adds
3473 the value in i to the value pointed to by j_ptr and returns the
3474 sum in the ax register. See \k{pushpop} for an explanation of
3475 \c{push} and \c{pop} and the use of context stacks.
3478 \S{stacksize} \i\c{%stacksize} Directive
3480 The \c{%stacksize} directive is used in conjunction with the
3481 \c{%arg} (see \k{arg}) and the \c{%local} (see \k{local}) directives.
3482 It tells NASM the default size to use for subsequent \c{%arg} and
3483 \c{%local} directives. The \c{%stacksize} directive takes one
3484 required argument which is one of \c{flat}, \c{flat64}, \c{large} or \c{small}.
3488 This form causes NASM to use stack-based parameter addressing
3489 relative to \c{ebp} and it assumes that a near form of call was used
3490 to get to this label (i.e. that \c{eip} is on the stack).
3492 \c %stacksize flat64
3494 This form causes NASM to use stack-based parameter addressing
3495 relative to \c{rbp} and it assumes that a near form of call was used
3496 to get to this label (i.e. that \c{rip} is on the stack).
3500 This form uses \c{bp} to do stack-based parameter addressing and
3501 assumes that a far form of call was used to get to this address
3502 (i.e. that \c{ip} and \c{cs} are on the stack).
3506 This form also uses \c{bp} to address stack parameters, but it is
3507 different from \c{large} because it also assumes that the old value
3508 of bp is pushed onto the stack (i.e. it expects an \c{ENTER}
3509 instruction). In other words, it expects that \c{bp}, \c{ip} and
3510 \c{cs} are on the top of the stack, underneath any local space which
3511 may have been allocated by \c{ENTER}. This form is probably most
3512 useful when used in combination with the \c{%local} directive
3516 \S{local} \i\c{%local} Directive
3518 The \c{%local} directive is used to simplify the use of local
3519 temporary stack variables allocated in a stack frame. Automatic
3520 local variables in C are an example of this kind of variable. The
3521 \c{%local} directive is most useful when used with the \c{%stacksize}
3522 (see \k{stacksize} and is also compatible with the \c{%arg} directive
3523 (see \k{arg}). It allows simplified reference to variables on the
3524 stack which have been allocated typically by using the \c{ENTER}
3526 \# (see \k{insENTER} for a description of that instruction).
3527 An example of its use is the following:
3531 \c %push mycontext ; save the current context
3532 \c %stacksize small ; tell NASM to use bp
3533 \c %assign %$localsize 0 ; see text for explanation
3534 \c %local old_ax:word, old_dx:word
3536 \c enter %$localsize,0 ; see text for explanation
3537 \c mov [old_ax],ax ; swap ax & bx
3538 \c mov [old_dx],dx ; and swap dx & cx
3543 \c leave ; restore old bp
3546 \c %pop ; restore original context
3548 The \c{%$localsize} variable is used internally by the
3549 \c{%local} directive and \e{must} be defined within the
3550 current context before the \c{%local} directive may be used.
3551 Failure to do so will result in one expression syntax error for
3552 each \c{%local} variable declared. It then may be used in
3553 the construction of an appropriately sized ENTER instruction
3554 as shown in the example.
3557 \H{pperror} Reporting \i{User-Defined Errors}: \i\c{%error}, \i\c{%warning}, \i\c{%fatal}
3559 The preprocessor directive \c{%error} will cause NASM to report an
3560 error if it occurs in assembled code. So if other users are going to
3561 try to assemble your source files, you can ensure that they define the
3562 right macros by means of code like this:
3567 \c ; do some different setup
3569 \c %error "Neither F1 nor F2 was defined."
3572 Then any user who fails to understand the way your code is supposed
3573 to be assembled will be quickly warned of their mistake, rather than
3574 having to wait until the program crashes on being run and then not
3575 knowing what went wrong.
3577 Similarly, \c{%warning} issues a warning, but allows assembly to continue:
3582 \c ; do some different setup
3584 \c %warning "Neither F1 nor F2 was defined, assuming F1."
3588 \c{%error} and \c{%warning} are issued only on the final assembly
3589 pass. This makes them safe to use in conjunction with tests that
3590 depend on symbol values.
3592 \c{%fatal} terminates assembly immediately, regardless of pass. This
3593 is useful when there is no point in continuing the assembly further,
3594 and doing so is likely just going to cause a spew of confusing error
3597 It is optional for the message string after \c{%error}, \c{%warning}
3598 or \c{%fatal} to be quoted. If it is \e{not}, then single-line macros
3599 are expanded in it, which can be used to display more information to
3600 the user. For example:
3603 \c %assign foo_over foo-64
3604 \c %error foo is foo_over bytes too large
3608 \H{otherpreproc} \i{Other Preprocessor Directives}
3610 NASM also has preprocessor directives which allow access to
3611 information from external sources. Currently they include:
3613 \b\c{%line} enables NASM to correctly handle the output of another
3614 preprocessor (see \k{line}).
3616 \b\c{%!} enables NASM to read in the value of an environment variable,
3617 which can then be used in your program (see \k{getenv}).
3619 \S{line} \i\c{%line} Directive
3621 The \c{%line} directive is used to notify NASM that the input line
3622 corresponds to a specific line number in another file. Typically
3623 this other file would be an original source file, with the current
3624 NASM input being the output of a pre-processor. The \c{%line}
3625 directive allows NASM to output messages which indicate the line
3626 number of the original source file, instead of the file that is being
3629 This preprocessor directive is not generally of use to programmers,
3630 by may be of interest to preprocessor authors. The usage of the
3631 \c{%line} preprocessor directive is as follows:
3633 \c %line nnn[+mmm] [filename]
3635 In this directive, \c{nnn} identifies the line of the original source
3636 file which this line corresponds to. \c{mmm} is an optional parameter
3637 which specifies a line increment value; each line of the input file
3638 read in is considered to correspond to \c{mmm} lines of the original
3639 source file. Finally, \c{filename} is an optional parameter which
3640 specifies the file name of the original source file.
3642 After reading a \c{%line} preprocessor directive, NASM will report
3643 all file name and line numbers relative to the values specified
3647 \S{getenv} \i\c{%!}\c{<env>}: Read an environment variable.
3649 The \c{%!<env>} directive makes it possible to read the value of an
3650 environment variable at assembly time. This could, for example, be used
3651 to store the contents of an environment variable into a string, which
3652 could be used at some other point in your code.
3654 For example, suppose that you have an environment variable \c{FOO}, and
3655 you want the contents of \c{FOO} to be embedded in your program. You
3656 could do that as follows:
3658 \c %defstr FOO %!FOO
3660 See \k{defstr} for notes on the \c{%defstr} directive.
3663 \H{stdmac} \i{Standard Macros}
3665 NASM defines a set of standard macros, which are already defined
3666 when it starts to process any source file. If you really need a
3667 program to be assembled with no pre-defined macros, you can use the
3668 \i\c{%clear} directive to empty the preprocessor of everything but
3669 context-local preprocessor variables and single-line macros.
3671 Most \i{user-level assembler directives} (see \k{directive}) are
3672 implemented as macros which invoke primitive directives; these are
3673 described in \k{directive}. The rest of the standard macro set is
3677 \S{stdmacver} \i{NASM Version} Macros
3679 The single-line macros \i\c{__NASM_MAJOR__}, \i\c{__NASM_MINOR__},
3680 \i\c{__NASM_SUBMINOR__} and \i\c{___NASM_PATCHLEVEL__} expand to the
3681 major, minor, subminor and patch level parts of the \i{version
3682 number of NASM} being used. So, under NASM 0.98.32p1 for
3683 example, \c{__NASM_MAJOR__} would be defined to be 0, \c{__NASM_MINOR__}
3684 would be defined as 98, \c{__NASM_SUBMINOR__} would be defined to 32,
3685 and \c{___NASM_PATCHLEVEL__} would be defined as 1.
3687 Additionally, the macro \i\c{__NASM_SNAPSHOT__} is defined for
3688 automatically generated snapshot releases \e{only}.
3691 \S{stdmacverid} \i\c{__NASM_VERSION_ID__}: \i{NASM Version ID}
3693 The single-line macro \c{__NASM_VERSION_ID__} expands to a dword integer
3694 representing the full version number of the version of nasm being used.
3695 The value is the equivalent to \c{__NASM_MAJOR__}, \c{__NASM_MINOR__},
3696 \c{__NASM_SUBMINOR__} and \c{___NASM_PATCHLEVEL__} concatenated to
3697 produce a single doubleword. Hence, for 0.98.32p1, the returned number
3698 would be equivalent to:
3706 Note that the above lines are generate exactly the same code, the second
3707 line is used just to give an indication of the order that the separate
3708 values will be present in memory.
3711 \S{stdmacverstr} \i\c{__NASM_VER__}: \i{NASM Version string}
3713 The single-line macro \c{__NASM_VER__} expands to a string which defines
3714 the version number of nasm being used. So, under NASM 0.98.32 for example,
3723 \S{fileline} \i\c{__FILE__} and \i\c{__LINE__}: File Name and Line Number
3725 Like the C preprocessor, NASM allows the user to find out the file
3726 name and line number containing the current instruction. The macro
3727 \c{__FILE__} expands to a string constant giving the name of the
3728 current input file (which may change through the course of assembly
3729 if \c{%include} directives are used), and \c{__LINE__} expands to a
3730 numeric constant giving the current line number in the input file.
3732 These macros could be used, for example, to communicate debugging
3733 information to a macro, since invoking \c{__LINE__} inside a macro
3734 definition (either single-line or multi-line) will return the line
3735 number of the macro \e{call}, rather than \e{definition}. So to
3736 determine where in a piece of code a crash is occurring, for
3737 example, one could write a routine \c{stillhere}, which is passed a
3738 line number in \c{EAX} and outputs something like `line 155: still
3739 here'. You could then write a macro
3741 \c %macro notdeadyet 0
3750 and then pepper your code with calls to \c{notdeadyet} until you
3751 find the crash point.
3754 \S{bitsm} \i\c{__BITS__}: Current BITS Mode
3756 The \c{__BITS__} standard macro is updated every time that the BITS mode is
3757 set using the \c{BITS XX} or \c{[BITS XX]} directive, where XX is a valid mode
3758 number of 16, 32 or 64. \c{__BITS__} receives the specified mode number and
3759 makes it globally available. This can be very useful for those who utilize
3760 mode-dependent macros.
3762 \S{ofmtm} \i\c{__OUTPUT_FORMAT__}: Current Output Format
3764 The \c{__OUTPUT_FORMAT__} standard macro holds the current Output Format,
3765 as given by the \c{-f} option or NASM's default. Type \c{nasm -hf} for a
3768 \c %ifidn __OUTPUT_FORMAT__, win32
3769 \c %define NEWLINE 13, 10
3770 \c %elifidn __OUTPUT_FORMAT__, elf32
3771 \c %define NEWLINE 10
3775 \S{datetime} Assembly Date and Time Macros
3777 NASM provides a variety of macros that represent the timestamp of the
3780 \b The \i\c{__DATE__} and \i\c{__TIME__} macros give the assembly date and
3781 time as strings, in ISO 8601 format (\c{"YYYY-MM-DD"} and \c{"HH:MM:SS"},
3784 \b The \i\c{__DATE_NUM__} and \i\c{__TIME_NUM__} macros give the assembly
3785 date and time in numeric form; in the format \c{YYYYMMDD} and
3786 \c{HHMMSS} respectively.
3788 \b The \i\c{__UTC_DATE__} and \i\c{__UTC_TIME__} macros give the assembly
3789 date and time in universal time (UTC) as strings, in ISO 8601 format
3790 (\c{"YYYY-MM-DD"} and \c{"HH:MM:SS"}, respectively.) If the host
3791 platform doesn't provide UTC time, these macros are undefined.
3793 \b The \i\c{__UTC_DATE_NUM__} and \i\c{__UTC_TIME_NUM__} macros give the
3794 assembly date and time universal time (UTC) in numeric form; in the
3795 format \c{YYYYMMDD} and \c{HHMMSS} respectively. If the
3796 host platform doesn't provide UTC time, these macros are
3799 \b The \c{__POSIX_TIME__} macro is defined as a number containing the
3800 number of seconds since the POSIX epoch, 1 January 1970 00:00:00 UTC;
3801 excluding any leap seconds. This is computed using UTC time if
3802 available on the host platform, otherwise it is computed using the
3803 local time as if it was UTC.
3805 All instances of time and date macros in the same assembly session
3806 produce consistent output. For example, in an assembly session
3807 started at 42 seconds after midnight on January 1, 2010 in Moscow
3808 (timezone UTC+3) these macros would have the following values,
3809 assuming, of course, a properly configured environment with a correct
3812 \c __DATE__ "2010-01-01"
3813 \c __TIME__ "00:00:42"
3814 \c __DATE_NUM__ 20100101
3815 \c __TIME_NUM__ 000042
3816 \c __UTC_DATE__ "2009-12-31"
3817 \c __UTC_TIME__ "21:00:42"
3818 \c __UTC_DATE_NUM__ 20091231
3819 \c __UTC_TIME_NUM__ 210042
3820 \c __POSIX_TIME__ 1262293242
3823 \S{use_def} \I\c{__USE_*__}\c{__USE_}\e{package}\c{__}: Package
3826 When a standard macro package (see \k{macropkg}) is included with the
3827 \c{%use} directive (see \k{use}), a single-line macro of the form
3828 \c{__USE_}\e{package}\c{__} is automatically defined. This allows
3829 testing if a particular package is invoked or not.
3831 For example, if the \c{altreg} package is included (see
3832 \k{pkg_altreg}), then the macro \c{__USE_ALTREG__} is defined.
3835 \S{pass_macro} \i\c{__PASS__}: Assembly Pass
3837 The macro \c{__PASS__} is defined to be \c{1} on preparatory passes,
3838 and \c{2} on the final pass. In preprocess-only mode, it is set to
3839 \c{3}, and when running only to generate dependencies (due to the
3840 \c{-M} or \c{-MG} option, see \k{opt-M}) it is set to \c{0}.
3842 \e{Avoid using this macro if at all possible. It is tremendously easy
3843 to generate very strange errors by misusing it, and the semantics may
3844 change in future versions of NASM.}
3847 \S{struc} \i\c{STRUC} and \i\c{ENDSTRUC}: \i{Declaring Structure} Data Types
3849 The core of NASM contains no intrinsic means of defining data
3850 structures; instead, the preprocessor is sufficiently powerful that
3851 data structures can be implemented as a set of macros. The macros
3852 \c{STRUC} and \c{ENDSTRUC} are used to define a structure data type.
3854 \c{STRUC} takes one or two parameters. The first parameter is the name
3855 of the data type. The second, optional parameter is the base offset of
3856 the structure. The name of the data type is defined as a symbol with
3857 the value of the base offset, and the name of the data type with the
3858 suffix \c{_size} appended to it is defined as an \c{EQU} giving the
3859 size of the structure. Once \c{STRUC} has been issued, you are
3860 defining the structure, and should define fields using the \c{RESB}
3861 family of pseudo-instructions, and then invoke \c{ENDSTRUC} to finish
3864 For example, to define a structure called \c{mytype} containing a
3865 longword, a word, a byte and a string of bytes, you might code
3876 The above code defines six symbols: \c{mt_long} as 0 (the offset
3877 from the beginning of a \c{mytype} structure to the longword field),
3878 \c{mt_word} as 4, \c{mt_byte} as 6, \c{mt_str} as 7, \c{mytype_size}
3879 as 39, and \c{mytype} itself as zero.
3881 The reason why the structure type name is defined at zero by default
3882 is a side effect of allowing structures to work with the local label
3883 mechanism: if your structure members tend to have the same names in
3884 more than one structure, you can define the above structure like this:
3895 This defines the offsets to the structure fields as \c{mytype.long},
3896 \c{mytype.word}, \c{mytype.byte} and \c{mytype.str}.
3898 NASM, since it has no \e{intrinsic} structure support, does not
3899 support any form of period notation to refer to the elements of a
3900 structure once you have one (except the above local-label notation),
3901 so code such as \c{mov ax,[mystruc.mt_word]} is not valid.
3902 \c{mt_word} is a constant just like any other constant, so the
3903 correct syntax is \c{mov ax,[mystruc+mt_word]} or \c{mov
3904 ax,[mystruc+mytype.word]}.
3906 Sometimes you only have the address of the structure displaced by an
3907 offset. For example, consider this standard stack frame setup:
3913 In this case, you could access an element by subtracting the offset:
3915 \c mov [ebp - 40 + mytype.word], ax
3917 However, if you do not want to repeat this offset, you can use -40 as
3920 \c struc mytype, -40
3922 And access an element this way:
3924 \c mov [ebp + mytype.word], ax
3927 \S{istruc} \i\c{ISTRUC}, \i\c{AT} and \i\c{IEND}: Declaring
3928 \i{Instances of Structures}
3930 Having defined a structure type, the next thing you typically want
3931 to do is to declare instances of that structure in your data
3932 segment. NASM provides an easy way to do this in the \c{ISTRUC}
3933 mechanism. To declare a structure of type \c{mytype} in a program,
3934 you code something like this:
3939 \c at mt_long, dd 123456
3940 \c at mt_word, dw 1024
3941 \c at mt_byte, db 'x'
3942 \c at mt_str, db 'hello, world', 13, 10, 0
3946 The function of the \c{AT} macro is to make use of the \c{TIMES}
3947 prefix to advance the assembly position to the correct point for the
3948 specified structure field, and then to declare the specified data.
3949 Therefore the structure fields must be declared in the same order as
3950 they were specified in the structure definition.
3952 If the data to go in a structure field requires more than one source
3953 line to specify, the remaining source lines can easily come after
3954 the \c{AT} line. For example:
3956 \c at mt_str, db 123,134,145,156,167,178,189
3959 Depending on personal taste, you can also omit the code part of the
3960 \c{AT} line completely, and start the structure field on the next
3964 \c db 'hello, world'
3968 \S{align} \i\c{ALIGN} and \i\c{ALIGNB}: Data Alignment
3970 The \c{ALIGN} and \c{ALIGNB} macros provides a convenient way to
3971 align code or data on a word, longword, paragraph or other boundary.
3972 (Some assemblers call this directive \i\c{EVEN}.) The syntax of the
3973 \c{ALIGN} and \c{ALIGNB} macros is
3975 \c align 4 ; align on 4-byte boundary
3976 \c align 16 ; align on 16-byte boundary
3977 \c align 8,db 0 ; pad with 0s rather than NOPs
3978 \c align 4,resb 1 ; align to 4 in the BSS
3979 \c alignb 4 ; equivalent to previous line
3981 Both macros require their first argument to be a power of two; they
3982 both compute the number of additional bytes required to bring the
3983 length of the current section up to a multiple of that power of two,
3984 and then apply the \c{TIMES} prefix to their second argument to
3985 perform the alignment.
3987 If the second argument is not specified, the default for \c{ALIGN}
3988 is \c{NOP}, and the default for \c{ALIGNB} is \c{RESB 1}. So if the
3989 second argument is specified, the two macros are equivalent.
3990 Normally, you can just use \c{ALIGN} in code and data sections and
3991 \c{ALIGNB} in BSS sections, and never need the second argument
3992 except for special purposes.
3994 \c{ALIGN} and \c{ALIGNB}, being simple macros, perform no error
3995 checking: they cannot warn you if their first argument fails to be a
3996 power of two, or if their second argument generates more than one
3997 byte of code. In each of these cases they will silently do the wrong
4000 \c{ALIGNB} (or \c{ALIGN} with a second argument of \c{RESB 1}) can
4001 be used within structure definitions:
4018 This will ensure that the structure members are sensibly aligned
4019 relative to the base of the structure.
4021 A final caveat: \c{ALIGN} and \c{ALIGNB} work relative to the
4022 beginning of the \e{section}, not the beginning of the address space
4023 in the final executable. Aligning to a 16-byte boundary when the
4024 section you're in is only guaranteed to be aligned to a 4-byte
4025 boundary, for example, is a waste of effort. Again, NASM does not
4026 check that the section's alignment characteristics are sensible for
4027 the use of \c{ALIGN} or \c{ALIGNB}.
4029 See also the \c{smartalign} standard macro package, \k{pkg_smartalign}.
4032 \C{macropkg} \i{Standard Macro Packages}
4034 The \i\c{%use} directive (see \k{use}) includes one of the standard
4035 macro packages included with the NASM distribution and compiled into
4036 the NASM binary. It operates like the \c{%include} directive (see
4037 \k{include}), but the included contents is provided by NASM itself.
4039 The names of standard macro packages are case insensitive, and can be
4043 \H{pkg_altreg} \i\c{altreg}: \i{Alternate Register Names}
4045 The \c{altreg} standard macro package provides alternate register
4046 names. It provides numeric register names for all registers (not just
4047 \c{R8}-\c{R15}), the Intel-defined aliases \c{R8L}-\c{R15L} for the
4048 low bytes of register (as opposed to the NASM/AMD standard names
4049 \c{R8B}-\c{R15B}), and the names \c{R0H}-\c{R3H} (by analogy with
4050 \c{R0L}-\c{R3L}) for \c{AH}, \c{CH}, \c{DH}, and \c{BH}.
4057 \c mov r0l,r3h ; mov al,bh
4063 \H{pkg_smartalign} \i\c{smartalign}\I{align, smart}: Smart \c{ALIGN} Macro
4065 The \c{smartalign} standard macro package provides for an \i\c{ALIGN}
4066 macro which is more powerful than the default (and
4067 backwards-compatible) one (see \k{align}). When the \c{smartalign}
4068 package is enabled, when \c{ALIGN} is used without a second argument,
4069 NASM will generate a sequence of instructions more efficient than a
4070 series of \c{NOP}. Furthermore, if the padding exceeds a specific
4071 threshold, then NASM will generate a jump over the entire padding
4074 The specific instructions generated can be controlled with the
4075 new \i\c{ALIGNMODE} macro. This macro takes two parameters: one mode,
4076 and an optional jump threshold override. The modes are as
4079 \b \c{generic}: Works on all x86 CPUs and should have reasonable
4080 performance. The default jump threshold is 8. This is the
4083 \b \c{nop}: Pad out with \c{NOP} instructions. The only difference
4084 compared to the standard \c{ALIGN} macro is that NASM can still jump
4085 over a large padding area. The default jump threshold is 16.
4087 \b \c{k7}: Optimize for the AMD K7 (Athlon/Althon XP). These
4088 instructions should still work on all x86 CPUs. The default jump
4091 \b \c{k8}: Optimize for the AMD K8 (Opteron/Althon 64). These
4092 instructions should still work on all x86 CPUs. The default jump
4095 \b \c{p6}: Optimize for Intel CPUs. This uses the long \c{NOP}
4096 instructions first introduced in Pentium Pro. This is incompatible
4097 with all CPUs of family 5 or lower, as well as some VIA CPUs and
4098 several virtualization solutions. The default jump threshold is 16.
4100 The macro \i\c{__ALIGNMODE__} is defined to contain the current
4101 alignment mode. A number of other macros beginning with \c{__ALIGN_}
4102 are used internally by this macro package.
4105 \C{directive} \i{Assembler Directives}
4107 NASM, though it attempts to avoid the bureaucracy of assemblers like
4108 MASM and TASM, is nevertheless forced to support a \e{few}
4109 directives. These are described in this chapter.
4111 NASM's directives come in two types: \I{user-level
4112 directives}\e{user-level} directives and \I{primitive
4113 directives}\e{primitive} directives. Typically, each directive has a
4114 user-level form and a primitive form. In almost all cases, we
4115 recommend that users use the user-level forms of the directives,
4116 which are implemented as macros which call the primitive forms.
4118 Primitive directives are enclosed in square brackets; user-level
4121 In addition to the universal directives described in this chapter,
4122 each object file format can optionally supply extra directives in
4123 order to control particular features of that file format. These
4124 \I{format-specific directives}\e{format-specific} directives are
4125 documented along with the formats that implement them, in \k{outfmt}.
4128 \H{bits} \i\c{BITS}: Specifying Target \i{Processor Mode}
4130 The \c{BITS} directive specifies whether NASM should generate code
4131 \I{16-bit mode, versus 32-bit mode}designed to run on a processor
4132 operating in 16-bit mode, 32-bit mode or 64-bit mode. The syntax is
4133 \c{BITS XX}, where XX is 16, 32 or 64.
4135 In most cases, you should not need to use \c{BITS} explicitly. The
4136 \c{aout}, \c{coff}, \c{elf}, \c{macho}, \c{win32} and \c{win64}
4137 object formats, which are designed for use in 32-bit or 64-bit
4138 operating systems, all cause NASM to select 32-bit or 64-bit mode,
4139 respectively, by default. The \c{obj} object format allows you
4140 to specify each segment you define as either \c{USE16} or \c{USE32},
4141 and NASM will set its operating mode accordingly, so the use of the
4142 \c{BITS} directive is once again unnecessary.
4144 The most likely reason for using the \c{BITS} directive is to write
4145 32-bit or 64-bit code in a flat binary file; this is because the \c{bin}
4146 output format defaults to 16-bit mode in anticipation of it being
4147 used most frequently to write DOS \c{.COM} programs, DOS \c{.SYS}
4148 device drivers and boot loader software.
4150 You do \e{not} need to specify \c{BITS 32} merely in order to use
4151 32-bit instructions in a 16-bit DOS program; if you do, the
4152 assembler will generate incorrect code because it will be writing
4153 code targeted at a 32-bit platform, to be run on a 16-bit one.
4155 When NASM is in \c{BITS 16} mode, instructions which use 32-bit
4156 data are prefixed with an 0x66 byte, and those referring to 32-bit
4157 addresses have an 0x67 prefix. In \c{BITS 32} mode, the reverse is
4158 true: 32-bit instructions require no prefixes, whereas instructions
4159 using 16-bit data need an 0x66 and those working on 16-bit addresses
4162 When NASM is in \c{BITS 64} mode, most instructions operate the same
4163 as they do for \c{BITS 32} mode. However, there are 8 more general and
4164 SSE registers, and 16-bit addressing is no longer supported.
4166 The default address size is 64 bits; 32-bit addressing can be selected
4167 with the 0x67 prefix. The default operand size is still 32 bits,
4168 however, and the 0x66 prefix selects 16-bit operand size. The \c{REX}
4169 prefix is used both to select 64-bit operand size, and to access the
4170 new registers. NASM automatically inserts REX prefixes when
4173 When the \c{REX} prefix is used, the processor does not know how to
4174 address the AH, BH, CH or DH (high 8-bit legacy) registers. Instead,
4175 it is possible to access the the low 8-bits of the SP, BP SI and DI
4176 registers as SPL, BPL, SIL and DIL, respectively; but only when the
4179 The \c{BITS} directive has an exactly equivalent primitive form,
4180 \c{[BITS 16]}, \c{[BITS 32]} and \c{[BITS 64]}. The user-level form is
4181 a macro which has no function other than to call the primitive form.
4183 Note that the space is neccessary, e.g. \c{BITS32} will \e{not} work!
4185 \S{USE16 & USE32} \i\c{USE16} & \i\c{USE32}: Aliases for BITS
4187 The `\c{USE16}' and `\c{USE32}' directives can be used in place of
4188 `\c{BITS 16}' and `\c{BITS 32}', for compatibility with other assemblers.
4191 \H{default} \i\c{DEFAULT}: Change the assembler defaults
4193 The \c{DEFAULT} directive changes the assembler defaults. Normally,
4194 NASM defaults to a mode where the programmer is expected to explicitly
4195 specify most features directly. However, this is occationally
4196 obnoxious, as the explicit form is pretty much the only one one wishes
4199 Currently, the only \c{DEFAULT} that is settable is whether or not
4200 registerless instructions in 64-bit mode are \c{RIP}-relative or not.
4201 By default, they are absolute unless overridden with the \i\c{REL}
4202 specifier (see \k{effaddr}). However, if \c{DEFAULT REL} is
4203 specified, \c{REL} is default, unless overridden with the \c{ABS}
4204 specifier, \e{except when used with an FS or GS segment override}.
4206 The special handling of \c{FS} and \c{GS} overrides are due to the
4207 fact that these registers are generally used as thread pointers or
4208 other special functions in 64-bit mode, and generating
4209 \c{RIP}-relative addresses would be extremely confusing.
4211 \c{DEFAULT REL} is disabled with \c{DEFAULT ABS}.
4213 \H{section} \i\c{SECTION} or \i\c{SEGMENT}: Changing and \i{Defining
4216 \I{changing sections}\I{switching between sections}The \c{SECTION}
4217 directive (\c{SEGMENT} is an exactly equivalent synonym) changes
4218 which section of the output file the code you write will be
4219 assembled into. In some object file formats, the number and names of
4220 sections are fixed; in others, the user may make up as many as they
4221 wish. Hence \c{SECTION} may sometimes give an error message, or may
4222 define a new section, if you try to switch to a section that does
4225 The Unix object formats, and the \c{bin} object format (but see
4226 \k{multisec}, all support
4227 the \i{standardized section names} \c{.text}, \c{.data} and \c{.bss}
4228 for the code, data and uninitialized-data sections. The \c{obj}
4229 format, by contrast, does not recognize these section names as being
4230 special, and indeed will strip off the leading period of any section
4234 \S{sectmac} The \i\c{__SECT__} Macro
4236 The \c{SECTION} directive is unusual in that its user-level form
4237 functions differently from its primitive form. The primitive form,
4238 \c{[SECTION xyz]}, simply switches the current target section to the
4239 one given. The user-level form, \c{SECTION xyz}, however, first
4240 defines the single-line macro \c{__SECT__} to be the primitive
4241 \c{[SECTION]} directive which it is about to issue, and then issues
4242 it. So the user-level directive
4246 expands to the two lines
4248 \c %define __SECT__ [SECTION .text]
4251 Users may find it useful to make use of this in their own macros.
4252 For example, the \c{writefile} macro defined in \k{mlmacgre} can be
4253 usefully rewritten in the following more sophisticated form:
4255 \c %macro writefile 2+
4265 \c mov cx,%%endstr-%%str
4272 This form of the macro, once passed a string to output, first
4273 switches temporarily to the data section of the file, using the
4274 primitive form of the \c{SECTION} directive so as not to modify
4275 \c{__SECT__}. It then declares its string in the data section, and
4276 then invokes \c{__SECT__} to switch back to \e{whichever} section
4277 the user was previously working in. It thus avoids the need, in the
4278 previous version of the macro, to include a \c{JMP} instruction to
4279 jump over the data, and also does not fail if, in a complicated
4280 \c{OBJ} format module, the user could potentially be assembling the
4281 code in any of several separate code sections.
4284 \H{absolute} \i\c{ABSOLUTE}: Defining Absolute Labels
4286 The \c{ABSOLUTE} directive can be thought of as an alternative form
4287 of \c{SECTION}: it causes the subsequent code to be directed at no
4288 physical section, but at the hypothetical section starting at the
4289 given absolute address. The only instructions you can use in this
4290 mode are the \c{RESB} family.
4292 \c{ABSOLUTE} is used as follows:
4300 This example describes a section of the PC BIOS data area, at
4301 segment address 0x40: the above code defines \c{kbuf_chr} to be
4302 0x1A, \c{kbuf_free} to be 0x1C, and \c{kbuf} to be 0x1E.
4304 The user-level form of \c{ABSOLUTE}, like that of \c{SECTION},
4305 redefines the \i\c{__SECT__} macro when it is invoked.
4307 \i\c{STRUC} and \i\c{ENDSTRUC} are defined as macros which use
4308 \c{ABSOLUTE} (and also \c{__SECT__}).
4310 \c{ABSOLUTE} doesn't have to take an absolute constant as an
4311 argument: it can take an expression (actually, a \i{critical
4312 expression}: see \k{crit}) and it can be a value in a segment. For
4313 example, a TSR can re-use its setup code as run-time BSS like this:
4315 \c org 100h ; it's a .COM program
4317 \c jmp setup ; setup code comes last
4319 \c ; the resident part of the TSR goes here
4321 \c ; now write the code that installs the TSR here
4325 \c runtimevar1 resw 1
4326 \c runtimevar2 resd 20
4330 This defines some variables `on top of' the setup code, so that
4331 after the setup has finished running, the space it took up can be
4332 re-used as data storage for the running TSR. The symbol `tsr_end'
4333 can be used to calculate the total size of the part of the TSR that
4334 needs to be made resident.
4337 \H{extern} \i\c{EXTERN}: \i{Importing Symbols} from Other Modules
4339 \c{EXTERN} is similar to the MASM directive \c{EXTRN} and the C
4340 keyword \c{extern}: it is used to declare a symbol which is not
4341 defined anywhere in the module being assembled, but is assumed to be
4342 defined in some other module and needs to be referred to by this
4343 one. Not every object-file format can support external variables:
4344 the \c{bin} format cannot.
4346 The \c{EXTERN} directive takes as many arguments as you like. Each
4347 argument is the name of a symbol:
4350 \c extern _sscanf,_fscanf
4352 Some object-file formats provide extra features to the \c{EXTERN}
4353 directive. In all cases, the extra features are used by suffixing a
4354 colon to the symbol name followed by object-format specific text.
4355 For example, the \c{obj} format allows you to declare that the
4356 default segment base of an external should be the group \c{dgroup}
4357 by means of the directive
4359 \c extern _variable:wrt dgroup
4361 The primitive form of \c{EXTERN} differs from the user-level form
4362 only in that it can take only one argument at a time: the support
4363 for multiple arguments is implemented at the preprocessor level.
4365 You can declare the same variable as \c{EXTERN} more than once: NASM
4366 will quietly ignore the second and later redeclarations. You can't
4367 declare a variable as \c{EXTERN} as well as something else, though.
4370 \H{global} \i\c{GLOBAL}: \i{Exporting Symbols} to Other Modules
4372 \c{GLOBAL} is the other end of \c{EXTERN}: if one module declares a
4373 symbol as \c{EXTERN} and refers to it, then in order to prevent
4374 linker errors, some other module must actually \e{define} the
4375 symbol and declare it as \c{GLOBAL}. Some assemblers use the name
4376 \i\c{PUBLIC} for this purpose.
4378 The \c{GLOBAL} directive applying to a symbol must appear \e{before}
4379 the definition of the symbol.
4381 \c{GLOBAL} uses the same syntax as \c{EXTERN}, except that it must
4382 refer to symbols which \e{are} defined in the same module as the
4383 \c{GLOBAL} directive. For example:
4389 \c{GLOBAL}, like \c{EXTERN}, allows object formats to define private
4390 extensions by means of a colon. The \c{elf} object format, for
4391 example, lets you specify whether global data items are functions or
4394 \c global hashlookup:function, hashtable:data
4396 Like \c{EXTERN}, the primitive form of \c{GLOBAL} differs from the
4397 user-level form only in that it can take only one argument at a
4401 \H{common} \i\c{COMMON}: Defining Common Data Areas
4403 The \c{COMMON} directive is used to declare \i\e{common variables}.
4404 A common variable is much like a global variable declared in the
4405 uninitialized data section, so that
4409 is similar in function to
4416 The difference is that if more than one module defines the same
4417 common variable, then at link time those variables will be
4418 \e{merged}, and references to \c{intvar} in all modules will point
4419 at the same piece of memory.
4421 Like \c{GLOBAL} and \c{EXTERN}, \c{COMMON} supports object-format
4422 specific extensions. For example, the \c{obj} format allows common
4423 variables to be NEAR or FAR, and the \c{elf} format allows you to
4424 specify the alignment requirements of a common variable:
4426 \c common commvar 4:near ; works in OBJ
4427 \c common intarray 100:4 ; works in ELF: 4 byte aligned
4429 Once again, like \c{EXTERN} and \c{GLOBAL}, the primitive form of
4430 \c{COMMON} differs from the user-level form only in that it can take
4431 only one argument at a time.
4434 \H{CPU} \i\c{CPU}: Defining CPU Dependencies
4436 The \i\c{CPU} directive restricts assembly to those instructions which
4437 are available on the specified CPU.
4441 \b\c{CPU 8086} Assemble only 8086 instruction set
4443 \b\c{CPU 186} Assemble instructions up to the 80186 instruction set
4445 \b\c{CPU 286} Assemble instructions up to the 286 instruction set
4447 \b\c{CPU 386} Assemble instructions up to the 386 instruction set
4449 \b\c{CPU 486} 486 instruction set
4451 \b\c{CPU 586} Pentium instruction set
4453 \b\c{CPU PENTIUM} Same as 586
4455 \b\c{CPU 686} P6 instruction set
4457 \b\c{CPU PPRO} Same as 686
4459 \b\c{CPU P2} Same as 686
4461 \b\c{CPU P3} Pentium III (Katmai) instruction sets
4463 \b\c{CPU KATMAI} Same as P3
4465 \b\c{CPU P4} Pentium 4 (Willamette) instruction set
4467 \b\c{CPU WILLAMETTE} Same as P4
4469 \b\c{CPU PRESCOTT} Prescott instruction set
4471 \b\c{CPU X64} x86-64 (x64/AMD64/Intel 64) instruction set
4473 \b\c{CPU IA64} IA64 CPU (in x86 mode) instruction set
4475 All options are case insensitive. All instructions will be selected
4476 only if they apply to the selected CPU or lower. By default, all
4477 instructions are available.
4480 \H{FLOAT} \i\c{FLOAT}: Handling of \I{floating-point, constants}floating-point constants
4482 By default, floating-point constants are rounded to nearest, and IEEE
4483 denormals are supported. The following options can be set to alter
4486 \b\c{FLOAT DAZ} Flush denormals to zero
4488 \b\c{FLOAT NODAZ} Do not flush denormals to zero (default)
4490 \b\c{FLOAT NEAR} Round to nearest (default)
4492 \b\c{FLOAT UP} Round up (toward +Infinity)
4494 \b\c{FLOAT DOWN} Round down (toward -Infinity)
4496 \b\c{FLOAT ZERO} Round toward zero
4498 \b\c{FLOAT DEFAULT} Restore default settings
4500 The standard macros \i\c{__FLOAT_DAZ__}, \i\c{__FLOAT_ROUND__}, and
4501 \i\c{__FLOAT__} contain the current state, as long as the programmer
4502 has avoided the use of the brackeded primitive form, (\c{[FLOAT]}).
4504 \c{__FLOAT__} contains the full set of floating-point settings; this
4505 value can be saved away and invoked later to restore the setting.
4508 \C{outfmt} \i{Output Formats}
4510 NASM is a portable assembler, designed to be able to compile on any
4511 ANSI C-supporting platform and produce output to run on a variety of
4512 Intel x86 operating systems. For this reason, it has a large number
4513 of available output formats, selected using the \i\c{-f} option on
4514 the NASM \i{command line}. Each of these formats, along with its
4515 extensions to the base NASM syntax, is detailed in this chapter.
4517 As stated in \k{opt-o}, NASM chooses a \i{default name} for your
4518 output file based on the input file name and the chosen output
4519 format. This will be generated by removing the \i{extension}
4520 (\c{.asm}, \c{.s}, or whatever you like to use) from the input file
4521 name, and substituting an extension defined by the output format.
4522 The extensions are given with each format below.
4525 \H{binfmt} \i\c{bin}: \i{Flat-Form Binary}\I{pure binary} Output
4527 The \c{bin} format does not produce object files: it generates
4528 nothing in the output file except the code you wrote. Such `pure
4529 binary' files are used by \i{MS-DOS}: \i\c{.COM} executables and
4530 \i\c{.SYS} device drivers are pure binary files. Pure binary output
4531 is also useful for \i{operating system} and \i{boot loader}
4534 The \c{bin} format supports \i{multiple section names}. For details of
4535 how NASM handles sections in the \c{bin} format, see \k{multisec}.
4537 Using the \c{bin} format puts NASM by default into 16-bit mode (see
4538 \k{bits}). In order to use \c{bin} to write 32-bit or 64-bit code,
4539 such as an OS kernel, you need to explicitly issue the \I\c{BITS}\c{BITS 32}
4540 or \I\c{BITS}\c{BITS 64} directive.
4542 \c{bin} has no default output file name extension: instead, it
4543 leaves your file name as it is once the original extension has been
4544 removed. Thus, the default is for NASM to assemble \c{binprog.asm}
4545 into a binary file called \c{binprog}.
4548 \S{org} \i\c{ORG}: Binary File \i{Program Origin}
4550 The \c{bin} format provides an additional directive to the list
4551 given in \k{directive}: \c{ORG}. The function of the \c{ORG}
4552 directive is to specify the origin address which NASM will assume
4553 the program begins at when it is loaded into memory.
4555 For example, the following code will generate the longword
4562 Unlike the \c{ORG} directive provided by MASM-compatible assemblers,
4563 which allows you to jump around in the object file and overwrite
4564 code you have already generated, NASM's \c{ORG} does exactly what
4565 the directive says: \e{origin}. Its sole function is to specify one
4566 offset which is added to all internal address references within the
4567 section; it does not permit any of the trickery that MASM's version
4568 does. See \k{proborg} for further comments.
4571 \S{binseg} \c{bin} Extensions to the \c{SECTION}
4572 Directive\I{SECTION, bin extensions to}
4574 The \c{bin} output format extends the \c{SECTION} (or \c{SEGMENT})
4575 directive to allow you to specify the alignment requirements of
4576 segments. This is done by appending the \i\c{ALIGN} qualifier to the
4577 end of the section-definition line. For example,
4579 \c section .data align=16
4581 switches to the section \c{.data} and also specifies that it must be
4582 aligned on a 16-byte boundary.
4584 The parameter to \c{ALIGN} specifies how many low bits of the
4585 section start address must be forced to zero. The alignment value
4586 given may be any power of two.\I{section alignment, in
4587 bin}\I{segment alignment, in bin}\I{alignment, in bin sections}
4590 \S{multisec} \i{Multisection}\I{bin, multisection} Support for the \c{bin} Format
4592 The \c{bin} format allows the use of multiple sections, of arbitrary names,
4593 besides the "known" \c{.text}, \c{.data}, and \c{.bss} names.
4595 \b Sections may be designated \i\c{progbits} or \i\c{nobits}. Default
4596 is \c{progbits} (except \c{.bss}, which defaults to \c{nobits},
4599 \b Sections can be aligned at a specified boundary following the previous
4600 section with \c{align=}, or at an arbitrary byte-granular position with
4603 \b Sections can be given a virtual start address, which will be used
4604 for the calculation of all memory references within that section
4607 \b Sections can be ordered using \i\c{follows=}\c{<section>} or
4608 \i\c{vfollows=}\c{<section>} as an alternative to specifying an explicit
4611 \b Arguments to \c{org}, \c{start}, \c{vstart}, and \c{align=} are
4612 critical expressions. See \k{crit}. E.g. \c{align=(1 << ALIGN_SHIFT)}
4613 - \c{ALIGN_SHIFT} must be defined before it is used here.
4615 \b Any code which comes before an explicit \c{SECTION} directive
4616 is directed by default into the \c{.text} section.
4618 \b If an \c{ORG} statement is not given, \c{ORG 0} is used
4621 \b The \c{.bss} section will be placed after the last \c{progbits}
4622 section, unless \c{start=}, \c{vstart=}, \c{follows=}, or \c{vfollows=}
4625 \b All sections are aligned on dword boundaries, unless a different
4626 alignment has been specified.
4628 \b Sections may not overlap.
4630 \b NASM creates the \c{section.<secname>.start} for each section,
4631 which may be used in your code.
4633 \S{map}\i{Map Files}
4635 Map files can be generated in \c{-f bin} format by means of the \c{[map]}
4636 option. Map types of \c{all} (default), \c{brief}, \c{sections}, \c{segments},
4637 or \c{symbols} may be specified. Output may be directed to \c{stdout}
4638 (default), \c{stderr}, or a specified file. E.g.
4639 \c{[map symbols myfile.map]}. No "user form" exists, the square
4640 brackets must be used.
4643 \H{ithfmt} \i\c{ith}: \i{Intel Hex} Output
4645 The \c{ith} file format produces Intel hex-format files. Just as the
4646 \c{bin} format, this is a flat memory image format with no support for
4647 relocation or linking. It is usually used with ROM programmers and
4650 All extensions supported by the \c{bin} file format is also supported by
4651 the \c{ith} file format.
4653 \c{ith} provides a default output file-name extension of \c{.ith}.
4656 \H{srecfmt} \i\c{srec}: \i{Motorola S-Records} Output
4658 The \c{srec} file format produces Motorola S-records files. Just as the
4659 \c{bin} format, this is a flat memory image format with no support for
4660 relocation or linking. It is usually used with ROM programmers and
4663 All extensions supported by the \c{bin} file format is also supported by
4664 the \c{srec} file format.
4666 \c{srec} provides a default output file-name extension of \c{.srec}.
4669 \H{objfmt} \i\c{obj}: \i{Microsoft OMF}\I{OMF} Object Files
4671 The \c{obj} file format (NASM calls it \c{obj} rather than \c{omf}
4672 for historical reasons) is the one produced by \i{MASM} and
4673 \i{TASM}, which is typically fed to 16-bit DOS linkers to produce
4674 \i\c{.EXE} files. It is also the format used by \i{OS/2}.
4676 \c{obj} provides a default output file-name extension of \c{.obj}.
4678 \c{obj} is not exclusively a 16-bit format, though: NASM has full
4679 support for the 32-bit extensions to the format. In particular,
4680 32-bit \c{obj} format files are used by \i{Borland's Win32
4681 compilers}, instead of using Microsoft's newer \i\c{win32} object
4684 The \c{obj} format does not define any special segment names: you
4685 can call your segments anything you like. Typical names for segments
4686 in \c{obj} format files are \c{CODE}, \c{DATA} and \c{BSS}.
4688 If your source file contains code before specifying an explicit
4689 \c{SEGMENT} directive, then NASM will invent its own segment called
4690 \i\c{__NASMDEFSEG} for you.
4692 When you define a segment in an \c{obj} file, NASM defines the
4693 segment name as a symbol as well, so that you can access the segment
4694 address of the segment. So, for example:
4703 \c mov ax,data ; get segment address of data
4704 \c mov ds,ax ; and move it into DS
4705 \c inc word [dvar] ; now this reference will work
4708 The \c{obj} format also enables the use of the \i\c{SEG} and
4709 \i\c{WRT} operators, so that you can write code which does things
4714 \c mov ax,seg foo ; get preferred segment of foo
4716 \c mov ax,data ; a different segment
4718 \c mov ax,[ds:foo] ; this accesses `foo'
4719 \c mov [es:foo wrt data],bx ; so does this
4722 \S{objseg} \c{obj} Extensions to the \c{SEGMENT}
4723 Directive\I{SEGMENT, obj extensions to}
4725 The \c{obj} output format extends the \c{SEGMENT} (or \c{SECTION})
4726 directive to allow you to specify various properties of the segment
4727 you are defining. This is done by appending extra qualifiers to the
4728 end of the segment-definition line. For example,
4730 \c segment code private align=16
4732 defines the segment \c{code}, but also declares it to be a private
4733 segment, and requires that the portion of it described in this code
4734 module must be aligned on a 16-byte boundary.
4736 The available qualifiers are:
4738 \b \i\c{PRIVATE}, \i\c{PUBLIC}, \i\c{COMMON} and \i\c{STACK} specify
4739 the combination characteristics of the segment. \c{PRIVATE} segments
4740 do not get combined with any others by the linker; \c{PUBLIC} and
4741 \c{STACK} segments get concatenated together at link time; and
4742 \c{COMMON} segments all get overlaid on top of each other rather
4743 than stuck end-to-end.
4745 \b \i\c{ALIGN} is used, as shown above, to specify how many low bits
4746 of the segment start address must be forced to zero. The alignment
4747 value given may be any power of two from 1 to 4096; in reality, the
4748 only values supported are 1, 2, 4, 16, 256 and 4096, so if 8 is
4749 specified it will be rounded up to 16, and 32, 64 and 128 will all
4750 be rounded up to 256, and so on. Note that alignment to 4096-byte
4751 boundaries is a \i{PharLap} extension to the format and may not be
4752 supported by all linkers.\I{section alignment, in OBJ}\I{segment
4753 alignment, in OBJ}\I{alignment, in OBJ sections}
4755 \b \i\c{CLASS} can be used to specify the segment class; this feature
4756 indicates to the linker that segments of the same class should be
4757 placed near each other in the output file. The class name can be any
4758 word, e.g. \c{CLASS=CODE}.
4760 \b \i\c{OVERLAY}, like \c{CLASS}, is specified with an arbitrary word
4761 as an argument, and provides overlay information to an
4762 overlay-capable linker.
4764 \b Segments can be declared as \i\c{USE16} or \i\c{USE32}, which has
4765 the effect of recording the choice in the object file and also
4766 ensuring that NASM's default assembly mode when assembling in that
4767 segment is 16-bit or 32-bit respectively.
4769 \b When writing \i{OS/2} object files, you should declare 32-bit
4770 segments as \i\c{FLAT}, which causes the default segment base for
4771 anything in the segment to be the special group \c{FLAT}, and also
4772 defines the group if it is not already defined.
4774 \b The \c{obj} file format also allows segments to be declared as
4775 having a pre-defined absolute segment address, although no linkers
4776 are currently known to make sensible use of this feature;
4777 nevertheless, NASM allows you to declare a segment such as
4778 \c{SEGMENT SCREEN ABSOLUTE=0xB800} if you need to. The \i\c{ABSOLUTE}
4779 and \c{ALIGN} keywords are mutually exclusive.
4781 NASM's default segment attributes are \c{PUBLIC}, \c{ALIGN=1}, no
4782 class, no overlay, and \c{USE16}.
4785 \S{group} \i\c{GROUP}: Defining Groups of Segments\I{segments, groups of}
4787 The \c{obj} format also allows segments to be grouped, so that a
4788 single segment register can be used to refer to all the segments in
4789 a group. NASM therefore supplies the \c{GROUP} directive, whereby
4798 \c ; some uninitialized data
4800 \c group dgroup data bss
4802 which will define a group called \c{dgroup} to contain the segments
4803 \c{data} and \c{bss}. Like \c{SEGMENT}, \c{GROUP} causes the group
4804 name to be defined as a symbol, so that you can refer to a variable
4805 \c{var} in the \c{data} segment as \c{var wrt data} or as \c{var wrt
4806 dgroup}, depending on which segment value is currently in your
4809 If you just refer to \c{var}, however, and \c{var} is declared in a
4810 segment which is part of a group, then NASM will default to giving
4811 you the offset of \c{var} from the beginning of the \e{group}, not
4812 the \e{segment}. Therefore \c{SEG var}, also, will return the group
4813 base rather than the segment base.
4815 NASM will allow a segment to be part of more than one group, but
4816 will generate a warning if you do this. Variables declared in a
4817 segment which is part of more than one group will default to being
4818 relative to the first group that was defined to contain the segment.
4820 A group does not have to contain any segments; you can still make
4821 \c{WRT} references to a group which does not contain the variable
4822 you are referring to. OS/2, for example, defines the special group
4823 \c{FLAT} with no segments in it.
4826 \S{uppercase} \i\c{UPPERCASE}: Disabling Case Sensitivity in Output
4828 Although NASM itself is \i{case sensitive}, some OMF linkers are
4829 not; therefore it can be useful for NASM to output single-case
4830 object files. The \c{UPPERCASE} format-specific directive causes all
4831 segment, group and symbol names that are written to the object file
4832 to be forced to upper case just before being written. Within a
4833 source file, NASM is still case-sensitive; but the object file can
4834 be written entirely in upper case if desired.
4836 \c{UPPERCASE} is used alone on a line; it requires no parameters.
4839 \S{import} \i\c{IMPORT}: Importing DLL Symbols\I{DLL symbols,
4840 importing}\I{symbols, importing from DLLs}
4842 The \c{IMPORT} format-specific directive defines a symbol to be
4843 imported from a DLL, for use if you are writing a DLL's \i{import
4844 library} in NASM. You still need to declare the symbol as \c{EXTERN}
4845 as well as using the \c{IMPORT} directive.
4847 The \c{IMPORT} directive takes two required parameters, separated by
4848 white space, which are (respectively) the name of the symbol you
4849 wish to import and the name of the library you wish to import it
4852 \c import WSAStartup wsock32.dll
4854 A third optional parameter gives the name by which the symbol is
4855 known in the library you are importing it from, in case this is not
4856 the same as the name you wish the symbol to be known by to your code
4857 once you have imported it. For example:
4859 \c import asyncsel wsock32.dll WSAAsyncSelect
4862 \S{export} \i\c{EXPORT}: Exporting DLL Symbols\I{DLL symbols,
4863 exporting}\I{symbols, exporting from DLLs}
4865 The \c{EXPORT} format-specific directive defines a global symbol to
4866 be exported as a DLL symbol, for use if you are writing a DLL in
4867 NASM. You still need to declare the symbol as \c{GLOBAL} as well as
4868 using the \c{EXPORT} directive.
4870 \c{EXPORT} takes one required parameter, which is the name of the
4871 symbol you wish to export, as it was defined in your source file. An
4872 optional second parameter (separated by white space from the first)
4873 gives the \e{external} name of the symbol: the name by which you
4874 wish the symbol to be known to programs using the DLL. If this name
4875 is the same as the internal name, you may leave the second parameter
4878 Further parameters can be given to define attributes of the exported
4879 symbol. These parameters, like the second, are separated by white
4880 space. If further parameters are given, the external name must also
4881 be specified, even if it is the same as the internal name. The
4882 available attributes are:
4884 \b \c{resident} indicates that the exported name is to be kept
4885 resident by the system loader. This is an optimisation for
4886 frequently used symbols imported by name.
4888 \b \c{nodata} indicates that the exported symbol is a function which
4889 does not make use of any initialized data.
4891 \b \c{parm=NNN}, where \c{NNN} is an integer, sets the number of
4892 parameter words for the case in which the symbol is a call gate
4893 between 32-bit and 16-bit segments.
4895 \b An attribute which is just a number indicates that the symbol
4896 should be exported with an identifying number (ordinal), and gives
4902 \c export myfunc TheRealMoreFormalLookingFunctionName
4903 \c export myfunc myfunc 1234 ; export by ordinal
4904 \c export myfunc myfunc resident parm=23 nodata
4907 \S{dotdotstart} \i\c{..start}: Defining the \i{Program Entry
4910 \c{OMF} linkers require exactly one of the object files being linked to
4911 define the program entry point, where execution will begin when the
4912 program is run. If the object file that defines the entry point is
4913 assembled using NASM, you specify the entry point by declaring the
4914 special symbol \c{..start} at the point where you wish execution to
4918 \S{objextern} \c{obj} Extensions to the \c{EXTERN}
4919 Directive\I{EXTERN, obj extensions to}
4921 If you declare an external symbol with the directive
4925 then references such as \c{mov ax,foo} will give you the offset of
4926 \c{foo} from its preferred segment base (as specified in whichever
4927 module \c{foo} is actually defined in). So to access the contents of
4928 \c{foo} you will usually need to do something like
4930 \c mov ax,seg foo ; get preferred segment base
4931 \c mov es,ax ; move it into ES
4932 \c mov ax,[es:foo] ; and use offset `foo' from it
4934 This is a little unwieldy, particularly if you know that an external
4935 is going to be accessible from a given segment or group, say
4936 \c{dgroup}. So if \c{DS} already contained \c{dgroup}, you could
4939 \c mov ax,[foo wrt dgroup]
4941 However, having to type this every time you want to access \c{foo}
4942 can be a pain; so NASM allows you to declare \c{foo} in the
4945 \c extern foo:wrt dgroup
4947 This form causes NASM to pretend that the preferred segment base of
4948 \c{foo} is in fact \c{dgroup}; so the expression \c{seg foo} will
4949 now return \c{dgroup}, and the expression \c{foo} is equivalent to
4952 This \I{default-WRT mechanism}default-\c{WRT} mechanism can be used
4953 to make externals appear to be relative to any group or segment in
4954 your program. It can also be applied to common variables: see
4958 \S{objcommon} \c{obj} Extensions to the \c{COMMON}
4959 Directive\I{COMMON, obj extensions to}
4961 The \c{obj} format allows common variables to be either near\I{near
4962 common variables} or far\I{far common variables}; NASM allows you to
4963 specify which your variables should be by the use of the syntax
4965 \c common nearvar 2:near ; `nearvar' is a near common
4966 \c common farvar 10:far ; and `farvar' is far
4968 Far common variables may be greater in size than 64Kb, and so the
4969 OMF specification says that they are declared as a number of
4970 \e{elements} of a given size. So a 10-byte far common variable could
4971 be declared as ten one-byte elements, five two-byte elements, two
4972 five-byte elements or one ten-byte element.
4974 Some \c{OMF} linkers require the \I{element size, in common
4975 variables}\I{common variables, element size}element size, as well as
4976 the variable size, to match when resolving common variables declared
4977 in more than one module. Therefore NASM must allow you to specify
4978 the element size on your far common variables. This is done by the
4981 \c common c_5by2 10:far 5 ; two five-byte elements
4982 \c common c_2by5 10:far 2 ; five two-byte elements
4984 If no element size is specified, the default is 1. Also, the \c{FAR}
4985 keyword is not required when an element size is specified, since
4986 only far commons may have element sizes at all. So the above
4987 declarations could equivalently be
4989 \c common c_5by2 10:5 ; two five-byte elements
4990 \c common c_2by5 10:2 ; five two-byte elements
4992 In addition to these extensions, the \c{COMMON} directive in \c{obj}
4993 also supports default-\c{WRT} specification like \c{EXTERN} does
4994 (explained in \k{objextern}). So you can also declare things like
4996 \c common foo 10:wrt dgroup
4997 \c common bar 16:far 2:wrt data
4998 \c common baz 24:wrt data:6
5001 \H{win32fmt} \i\c{win32}: Microsoft Win32 Object Files
5003 The \c{win32} output format generates Microsoft Win32 object files,
5004 suitable for passing to Microsoft linkers such as \i{Visual C++}.
5005 Note that Borland Win32 compilers do not use this format, but use
5006 \c{obj} instead (see \k{objfmt}).
5008 \c{win32} provides a default output file-name extension of \c{.obj}.
5010 Note that although Microsoft say that Win32 object files follow the
5011 \c{COFF} (Common Object File Format) standard, the object files produced
5012 by Microsoft Win32 compilers are not compatible with COFF linkers
5013 such as DJGPP's, and vice versa. This is due to a difference of
5014 opinion over the precise semantics of PC-relative relocations. To
5015 produce COFF files suitable for DJGPP, use NASM's \c{coff} output
5016 format; conversely, the \c{coff} format does not produce object
5017 files that Win32 linkers can generate correct output from.
5020 \S{win32sect} \c{win32} Extensions to the \c{SECTION}
5021 Directive\I{SECTION, win32 extensions to}
5023 Like the \c{obj} format, \c{win32} allows you to specify additional
5024 information on the \c{SECTION} directive line, to control the type
5025 and properties of sections you declare. Section types and properties
5026 are generated automatically by NASM for the \i{standard section names}
5027 \c{.text}, \c{.data} and \c{.bss}, but may still be overridden by
5030 The available qualifiers are:
5032 \b \c{code}, or equivalently \c{text}, defines the section to be a
5033 code section. This marks the section as readable and executable, but
5034 not writable, and also indicates to the linker that the type of the
5037 \b \c{data} and \c{bss} define the section to be a data section,
5038 analogously to \c{code}. Data sections are marked as readable and
5039 writable, but not executable. \c{data} declares an initialized data
5040 section, whereas \c{bss} declares an uninitialized data section.
5042 \b \c{rdata} declares an initialized data section that is readable
5043 but not writable. Microsoft compilers use this section to place
5046 \b \c{info} defines the section to be an \i{informational section},
5047 which is not included in the executable file by the linker, but may
5048 (for example) pass information \e{to} the linker. For example,
5049 declaring an \c{info}-type section called \i\c{.drectve} causes the
5050 linker to interpret the contents of the section as command-line
5053 \b \c{align=}, used with a trailing number as in \c{obj}, gives the
5054 \I{section alignment, in win32}\I{alignment, in win32
5055 sections}alignment requirements of the section. The maximum you may
5056 specify is 64: the Win32 object file format contains no means to
5057 request a greater section alignment than this. If alignment is not
5058 explicitly specified, the defaults are 16-byte alignment for code
5059 sections, 8-byte alignment for rdata sections and 4-byte alignment
5060 for data (and BSS) sections.
5061 Informational sections get a default alignment of 1 byte (no
5062 alignment), though the value does not matter.
5064 The defaults assumed by NASM if you do not specify the above
5067 \c section .text code align=16
5068 \c section .data data align=4
5069 \c section .rdata rdata align=8
5070 \c section .bss bss align=4
5072 Any other section name is treated by default like \c{.text}.
5074 \S{win32safeseh} \c{win32}: Safe Structured Exception Handling
5076 Among other improvements in Windows XP SP2 and Windows Server 2003
5077 Microsoft has introduced concept of "safe structured exception
5078 handling." General idea is to collect handlers' entry points in
5079 designated read-only table and have alleged entry point verified
5080 against this table prior exception control is passed to the handler. In
5081 order for an executable module to be equipped with such "safe exception
5082 handler table," all object modules on linker command line has to comply
5083 with certain criteria. If one single module among them does not, then
5084 the table in question is omitted and above mentioned run-time checks
5085 will not be performed for application in question. Table omission is by
5086 default silent and therefore can be easily overlooked. One can instruct
5087 linker to refuse to produce binary without such table by passing
5088 \c{/safeseh} command line option.
5090 Without regard to this run-time check merits it's natural to expect
5091 NASM to be capable of generating modules suitable for \c{/safeseh}
5092 linking. From developer's viewpoint the problem is two-fold:
5094 \b how to adapt modules not deploying exception handlers of their own;
5096 \b how to adapt/develop modules utilizing custom exception handling;
5098 Former can be easily achieved with any NASM version by adding following
5099 line to source code:
5103 As of version 2.03 NASM adds this absolute symbol automatically. If
5104 it's not already present to be precise. I.e. if for whatever reason
5105 developer would choose to assign another value in source file, it would
5106 still be perfectly possible.
5108 Registering custom exception handler on the other hand requires certain
5109 "magic." As of version 2.03 additional directive is implemented,
5110 \c{safeseh}, which instructs the assembler to produce appropriately
5111 formatted input data for above mentioned "safe exception handler
5112 table." Its typical use would be:
5115 \c extern _MessageBoxA@16
5116 \c %if __NASM_VERSION_ID__ >= 0x02030000
5117 \c safeseh handler ; register handler as "safe handler"
5120 \c push DWORD 1 ; MB_OKCANCEL
5121 \c push DWORD caption
5124 \c call _MessageBoxA@16
5125 \c sub eax,1 ; incidentally suits as return value
5126 \c ; for exception handler
5130 \c push DWORD handler
5131 \c push DWORD [fs:0]
5132 \c mov DWORD [fs:0],esp ; engage exception handler
5134 \c mov eax,DWORD[eax] ; cause exception
5135 \c pop DWORD [fs:0] ; disengage exception handler
5138 \c text: db 'OK to rethrow, CANCEL to generate core dump',0
5139 \c caption:db 'SEGV',0
5141 \c section .drectve info
5142 \c db '/defaultlib:user32.lib /defaultlib:msvcrt.lib '
5144 As you might imagine, it's perfectly possible to produce .exe binary
5145 with "safe exception handler table" and yet engage unregistered
5146 exception handler. Indeed, handler is engaged by simply manipulating
5147 \c{[fs:0]} location at run-time, something linker has no power over,
5148 run-time that is. It should be explicitly mentioned that such failure
5149 to register handler's entry point with \c{safeseh} directive has
5150 undesired side effect at run-time. If exception is raised and
5151 unregistered handler is to be executed, the application is abruptly
5152 terminated without any notification whatsoever. One can argue that
5153 system could at least have logged some kind "non-safe exception
5154 handler in x.exe at address n" message in event log, but no, literally
5155 no notification is provided and user is left with no clue on what
5156 caused application failure.
5158 Finally, all mentions of linker in this paragraph refer to Microsoft
5159 linker version 7.x and later. Presence of \c{@feat.00} symbol and input
5160 data for "safe exception handler table" causes no backward
5161 incompatibilities and "safeseh" modules generated by NASM 2.03 and
5162 later can still be linked by earlier versions or non-Microsoft linkers.
5165 \H{win64fmt} \i\c{win64}: Microsoft Win64 Object Files
5167 The \c{win64} output format generates Microsoft Win64 object files,
5168 which is nearly 100% identical to the \c{win32} object format (\k{win32fmt})
5169 with the exception that it is meant to target 64-bit code and the x86-64
5170 platform altogether. This object file is used exactly the same as the \c{win32}
5171 object format (\k{win32fmt}), in NASM, with regard to this exception.
5173 \S{win64pic} \c{win64}: Writing Position-Independent Code
5175 While \c{REL} takes good care of RIP-relative addressing, there is one
5176 aspect that is easy to overlook for a Win64 programmer: indirect
5177 references. Consider a switch dispatch table:
5179 \c jmp QWORD[dsptch+rax*8]
5185 Even novice Win64 assembler programmer will soon realize that the code
5186 is not 64-bit savvy. Most notably linker will refuse to link it with
5187 "\c{'ADDR32' relocation to '.text' invalid without
5188 /LARGEADDRESSAWARE:NO}". So [s]he will have to split jmp instruction as
5191 \c lea rbx,[rel dsptch]
5192 \c jmp QWORD[rbx+rax*8]
5194 What happens behind the scene is that effective address in \c{lea} is
5195 encoded relative to instruction pointer, or in perfectly
5196 position-independent manner. But this is only part of the problem!
5197 Trouble is that in .dll context \c{caseN} relocations will make their
5198 way to the final module and might have to be adjusted at .dll load
5199 time. To be specific when it can't be loaded at preferred address. And
5200 when this occurs, pages with such relocations will be rendered private
5201 to current process, which kind of undermines the idea of sharing .dll.
5202 But no worry, it's trivial to fix:
5204 \c lea rbx,[rel dsptch]
5205 \c add rbx,QWORD[rbx+rax*8]
5208 \c dsptch: dq case0-dsptch
5212 NASM version 2.03 and later provides another alternative, \c{wrt
5213 ..imagebase} operator, which returns offset from base address of the
5214 current image, be it .exe or .dll module, therefore the name. For those
5215 acquainted with PE-COFF format base address denotes start of
5216 \c{IMAGE_DOS_HEADER} structure. Here is how to implement switch with
5217 these image-relative references:
5219 \c lea rbx,[rel dsptch]
5220 \c mov eax,DWORD[rbx+rax*4]
5221 \c sub rbx,dsptch wrt ..imagebase
5225 \c dsptch: dd case0 wrt ..imagebase
5226 \c dd case1 wrt ..imagebase
5228 One can argue that the operator is redundant. Indeed, snippet before
5229 last works just fine with any NASM version and is not even Windows
5230 specific... The real reason for implementing \c{wrt ..imagebase} will
5231 become apparent in next paragraph.
5233 It should be noted that \c{wrt ..imagebase} is defined as 32-bit
5236 \c dd label wrt ..imagebase ; ok
5237 \c dq label wrt ..imagebase ; bad
5238 \c mov eax,label wrt ..imagebase ; ok
5239 \c mov rax,label wrt ..imagebase ; bad
5241 \S{win64seh} \c{win64}: Structured Exception Handling
5243 Structured exception handing in Win64 is completely different matter
5244 from Win32. Upon exception program counter value is noted, and
5245 linker-generated table comprising start and end addresses of all the
5246 functions [in given executable module] is traversed and compared to the
5247 saved program counter. Thus so called \c{UNWIND_INFO} structure is
5248 identified. If it's not found, then offending subroutine is assumed to
5249 be "leaf" and just mentioned lookup procedure is attempted for its
5250 caller. In Win64 leaf function is such function that does not call any
5251 other function \e{nor} modifies any Win64 non-volatile registers,
5252 including stack pointer. The latter ensures that it's possible to
5253 identify leaf function's caller by simply pulling the value from the
5256 While majority of subroutines written in assembler are not calling any
5257 other function, requirement for non-volatile registers' immutability
5258 leaves developer with not more than 7 registers and no stack frame,
5259 which is not necessarily what [s]he counted with. Customarily one would
5260 meet the requirement by saving non-volatile registers on stack and
5261 restoring them upon return, so what can go wrong? If [and only if] an
5262 exception is raised at run-time and no \c{UNWIND_INFO} structure is
5263 associated with such "leaf" function, the stack unwind procedure will
5264 expect to find caller's return address on the top of stack immediately
5265 followed by its frame. Given that developer pushed caller's
5266 non-volatile registers on stack, would the value on top point at some
5267 code segment or even addressable space? Well, developer can attempt
5268 copying caller's return address to the top of stack and this would
5269 actually work in some very specific circumstances. But unless developer
5270 can guarantee that these circumstances are always met, it's more
5271 appropriate to assume worst case scenario, i.e. stack unwind procedure
5272 going berserk. Relevant question is what happens then? Application is
5273 abruptly terminated without any notification whatsoever. Just like in
5274 Win32 case, one can argue that system could at least have logged
5275 "unwind procedure went berserk in x.exe at address n" in event log, but
5276 no, no trace of failure is left.
5278 Now, when we understand significance of the \c{UNWIND_INFO} structure,
5279 let's discuss what's in it and/or how it's processed. First of all it
5280 is checked for presence of reference to custom language-specific
5281 exception handler. If there is one, then it's invoked. Depending on the
5282 return value, execution flow is resumed (exception is said to be
5283 "handled"), \e{or} rest of \c{UNWIND_INFO} structure is processed as
5284 following. Beside optional reference to custom handler, it carries
5285 information about current callee's stack frame and where non-volatile
5286 registers are saved. Information is detailed enough to be able to
5287 reconstruct contents of caller's non-volatile registers upon call to
5288 current callee. And so caller's context is reconstructed, and then
5289 unwind procedure is repeated, i.e. another \c{UNWIND_INFO} structure is
5290 associated, this time, with caller's instruction pointer, which is then
5291 checked for presence of reference to language-specific handler, etc.
5292 The procedure is recursively repeated till exception is handled. As
5293 last resort system "handles" it by generating memory core dump and
5294 terminating the application.
5296 As for the moment of this writing NASM unfortunately does not
5297 facilitate generation of above mentioned detailed information about
5298 stack frame layout. But as of version 2.03 it implements building
5299 blocks for generating structures involved in stack unwinding. As
5300 simplest example, here is how to deploy custom exception handler for
5305 \c extern MessageBoxA
5311 \c mov r9,1 ; MB_OKCANCEL
5313 \c sub eax,1 ; incidentally suits as return value
5314 \c ; for exception handler
5320 \c mov rax,QWORD[rax] ; cause exception
5323 \c text: db 'OK to rethrow, CANCEL to generate core dump',0
5324 \c caption:db 'SEGV',0
5326 \c section .pdata rdata align=4
5327 \c dd main wrt ..imagebase
5328 \c dd main_end wrt ..imagebase
5329 \c dd xmain wrt ..imagebase
5330 \c section .xdata rdata align=8
5331 \c xmain: db 9,0,0,0
5332 \c dd handler wrt ..imagebase
5333 \c section .drectve info
5334 \c db '/defaultlib:user32.lib /defaultlib:msvcrt.lib '
5336 What you see in \c{.pdata} section is element of the "table comprising
5337 start and end addresses of function" along with reference to associated
5338 \c{UNWIND_INFO} structure. And what you see in \c{.xdata} section is
5339 \c{UNWIND_INFO} structure describing function with no frame, but with
5340 designated exception handler. References are \e{required} to be
5341 image-relative (which is the real reason for implementing \c{wrt
5342 ..imagebase} operator). It should be noted that \c{rdata align=n}, as
5343 well as \c{wrt ..imagebase}, are optional in these two segments'
5344 contexts, i.e. can be omitted. Latter means that \e{all} 32-bit
5345 references, not only above listed required ones, placed into these two
5346 segments turn out image-relative. Why is it important to understand?
5347 Developer is allowed to append handler-specific data to \c{UNWIND_INFO}
5348 structure, and if [s]he adds a 32-bit reference, then [s]he will have
5349 to remember to adjust its value to obtain the real pointer.
5351 As already mentioned, in Win64 terms leaf function is one that does not
5352 call any other function \e{nor} modifies any non-volatile register,
5353 including stack pointer. But it's not uncommon that assembler
5354 programmer plans to utilize every single register and sometimes even
5355 have variable stack frame. Is there anything one can do with bare
5356 building blocks? I.e. besides manually composing fully-fledged
5357 \c{UNWIND_INFO} structure, which would surely be considered
5358 error-prone? Yes, there is. Recall that exception handler is called
5359 first, before stack layout is analyzed. As it turned out, it's
5360 perfectly possible to manipulate current callee's context in custom
5361 handler in manner that permits further stack unwinding. General idea is
5362 that handler would not actually "handle" the exception, but instead
5363 restore callee's context, as it was at its entry point and thus mimic
5364 leaf function. In other words, handler would simply undertake part of
5365 unwinding procedure. Consider following example:
5368 \c mov rax,rsp ; copy rsp to volatile register
5369 \c push r15 ; save non-volatile registers
5372 \c mov r11,rsp ; prepare variable stack frame
5375 \c mov QWORD[r11],rax ; check for exceptions
5376 \c mov rsp,r11 ; allocate stack frame
5377 \c mov QWORD[rsp],rax ; save original rsp value
5380 \c mov r11,QWORD[rsp] ; pull original rsp value
5381 \c mov rbp,QWORD[r11-24]
5382 \c mov rbx,QWORD[r11-16]
5383 \c mov r15,QWORD[r11-8]
5384 \c mov rsp,r11 ; destroy frame
5387 The keyword is that up to \c{magic_point} original \c{rsp} value
5388 remains in chosen volatile register and no non-volatile register,
5389 except for \c{rsp}, is modified. While past \c{magic_point} \c{rsp}
5390 remains constant till the very end of the \c{function}. In this case
5391 custom language-specific exception handler would look like this:
5393 \c EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
5394 \c CONTEXT *context,DISPATCHER_CONTEXT *disp)
5396 \c if (context->Rip<(ULONG64)magic_point)
5397 \c rsp = (ULONG64 *)context->Rax;
5399 \c { rsp = ((ULONG64 **)context->Rsp)[0];
5400 \c context->Rbp = rsp[-3];
5401 \c context->Rbx = rsp[-2];
5402 \c context->R15 = rsp[-1];
5404 \c context->Rsp = (ULONG64)rsp;
5406 \c memcpy (disp->ContextRecord,context,sizeof(CONTEXT));
5407 \c RtlVirtualUnwind(UNW_FLAG_NHANDLER,disp->ImageBase,
5408 \c dips->ControlPc,disp->FunctionEntry,disp->ContextRecord,
5409 \c &disp->HandlerData,&disp->EstablisherFrame,NULL);
5410 \c return ExceptionContinueSearch;
5413 As custom handler mimics leaf function, corresponding \c{UNWIND_INFO}
5414 structure does not have to contain any information about stack frame
5417 \H{cofffmt} \i\c{coff}: \i{Common Object File Format}
5419 The \c{coff} output type produces \c{COFF} object files suitable for
5420 linking with the \i{DJGPP} linker.
5422 \c{coff} provides a default output file-name extension of \c{.o}.
5424 The \c{coff} format supports the same extensions to the \c{SECTION}
5425 directive as \c{win32} does, except that the \c{align} qualifier and
5426 the \c{info} section type are not supported.
5428 \H{machofmt} \I{Mach-O}\i\c{macho32} and \i\c{macho64}: \i{Mach Object File Format}
5430 The \c{macho32} and \c{macho64} output formts produces \c{Mach-O}
5431 object files suitable for linking with the \i{MacOS X} linker.
5432 \i\c{macho} is a synonym for \c{macho32}.
5434 \c{macho} provides a default output file-name extension of \c{.o}.
5436 \H{elffmt} \i\c{elf32} and \i\c{elf64}: \I{ELF}\I{linux, elf}\i{Executable and Linkable
5437 Format} Object Files
5439 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},
5440 including \i{Solaris x86}, \i{UnixWare} and \i{SCO Unix}. \c{elf}
5441 provides a default output file-name extension of \c{.o}.
5442 \c{elf} is a synonym for \c{elf32}.
5444 \S{abisect} ELF specific directive \i\c{osabi}
5446 The ELF header specifies the application binary interface for the target operating system (OSABI).
5447 This field can be set by using the \c{osabi} directive with the numeric value (0-255) of the target
5448 system. If this directive is not used, the default value will be "UNIX System V ABI" (0) which will work on
5449 most systems which support ELF.
5451 \S{elfsect} \c{elf} Extensions to the \c{SECTION}
5452 Directive\I{SECTION, elf extensions to}
5454 Like the \c{obj} format, \c{elf} allows you to specify additional
5455 information on the \c{SECTION} directive line, to control the type
5456 and properties of sections you declare. Section types and properties
5457 are generated automatically by NASM for the \i{standard section
5458 names}, but may still be
5459 overridden by these qualifiers.
5461 The available qualifiers are:
5463 \b \i\c{alloc} defines the section to be one which is loaded into
5464 memory when the program is run. \i\c{noalloc} defines it to be one
5465 which is not, such as an informational or comment section.
5467 \b \i\c{exec} defines the section to be one which should have execute
5468 permission when the program is run. \i\c{noexec} defines it as one
5471 \b \i\c{write} defines the section to be one which should be writable
5472 when the program is run. \i\c{nowrite} defines it as one which should
5475 \b \i\c{progbits} defines the section to be one with explicit contents
5476 stored in the object file: an ordinary code or data section, for
5477 example, \i\c{nobits} defines the section to be one with no explicit
5478 contents given, such as a BSS section.
5480 \b \c{align=}, used with a trailing number as in \c{obj}, gives the
5481 \I{section alignment, in elf}\I{alignment, in elf sections}alignment
5482 requirements of the section.
5484 \b \i\c{tls} defines the section to be one which contains
5485 thread local variables.
5487 The defaults assumed by NASM if you do not specify the above
5490 \I\c{.text} \I\c{.rodata} \I\c{.lrodata} \I\c{.data} \I\c{.ldata}
5491 \I\c{.bss} \I\c{.lbss} \I\c{.tdata} \I\c{.tbss} \I\c\{.comment}
5493 \c section .text progbits alloc exec nowrite align=16
5494 \c section .rodata progbits alloc noexec nowrite align=4
5495 \c section .lrodata progbits alloc noexec nowrite align=4
5496 \c section .data progbits alloc noexec write align=4
5497 \c section .ldata progbits alloc noexec write align=4
5498 \c section .bss nobits alloc noexec write align=4
5499 \c section .lbss nobits alloc noexec write align=4
5500 \c section .tdata progbits alloc noexec write align=4 tls
5501 \c section .tbss nobits alloc noexec write align=4 tls
5502 \c section .comment progbits noalloc noexec nowrite align=1
5503 \c section other progbits alloc noexec nowrite align=1
5505 (Any section name other than those in the above table
5506 is treated by default like \c{other} in the above table.
5507 Please note that section names are case sensitive.)
5510 \S{elfwrt} \i{Position-Independent Code}\I{PIC}: \c{elf} Special
5511 Symbols and \i\c{WRT}
5513 The \c{ELF} specification contains enough features to allow
5514 position-independent code (PIC) to be written, which makes \i{ELF
5515 shared libraries} very flexible. However, it also means NASM has to
5516 be able to generate a variety of ELF specific relocation types in ELF
5517 object files, if it is to be an assembler which can write PIC.
5519 Since \c{ELF} does not support segment-base references, the \c{WRT}
5520 operator is not used for its normal purpose; therefore NASM's
5521 \c{elf} output format makes use of \c{WRT} for a different purpose,
5522 namely the PIC-specific \I{relocations, PIC-specific}relocation
5525 \c{elf} defines five special symbols which you can use as the
5526 right-hand side of the \c{WRT} operator to obtain PIC relocation
5527 types. They are \i\c{..gotpc}, \i\c{..gotoff}, \i\c{..got},
5528 \i\c{..plt} and \i\c{..sym}. Their functions are summarized here:
5530 \b Referring to the symbol marking the global offset table base
5531 using \c{wrt ..gotpc} will end up giving the distance from the
5532 beginning of the current section to the global offset table.
5533 (\i\c{_GLOBAL_OFFSET_TABLE_} is the standard symbol name used to
5534 refer to the \i{GOT}.) So you would then need to add \i\c{$$} to the
5535 result to get the real address of the GOT.
5537 \b Referring to a location in one of your own sections using \c{wrt
5538 ..gotoff} will give the distance from the beginning of the GOT to
5539 the specified location, so that adding on the address of the GOT
5540 would give the real address of the location you wanted.
5542 \b Referring to an external or global symbol using \c{wrt ..got}
5543 causes the linker to build an entry \e{in} the GOT containing the
5544 address of the symbol, and the reference gives the distance from the
5545 beginning of the GOT to the entry; so you can add on the address of
5546 the GOT, load from the resulting address, and end up with the
5547 address of the symbol.
5549 \b Referring to a procedure name using \c{wrt ..plt} causes the
5550 linker to build a \i{procedure linkage table} entry for the symbol,
5551 and the reference gives the address of the \i{PLT} entry. You can
5552 only use this in contexts which would generate a PC-relative
5553 relocation normally (i.e. as the destination for \c{CALL} or
5554 \c{JMP}), since ELF contains no relocation type to refer to PLT
5557 \b Referring to a symbol name using \c{wrt ..sym} causes NASM to
5558 write an ordinary relocation, but instead of making the relocation
5559 relative to the start of the section and then adding on the offset
5560 to the symbol, it will write a relocation record aimed directly at
5561 the symbol in question. The distinction is a necessary one due to a
5562 peculiarity of the dynamic linker.
5564 A fuller explanation of how to use these relocation types to write
5565 shared libraries entirely in NASM is given in \k{picdll}.
5567 \S{elftls} \i{Thread Local Storage}\I{TLS}: \c{elf} Special
5568 Symbols and \i\c{WRT}
5570 \b In ELF32 mode, referring to an external or global symbol using
5571 \c{wrt ..tlsie} \I\c{..tlsie}
5572 causes the linker to build an entry \e{in} the GOT containing the
5573 offset of the symbol within the TLS block, so you can access the value
5574 of the symbol with code such as:
5576 \c mov eax,[tid wrt ..tlsie]
5580 \b In ELF64 mode, referring to an external or global symbol using
5581 \c{wrt ..gottpoff} \I\c{..gottpoff}
5582 causes the linker to build an entry \e{in} the GOT containing the
5583 offset of the symbol within the TLS block, so you can access the value
5584 of the symbol with code such as:
5586 \c mov rax,[rel tid wrt ..gottpoff]
5590 \S{elfglob} \c{elf} Extensions to the \c{GLOBAL} Directive\I{GLOBAL,
5591 elf extensions to}\I{GLOBAL, aoutb extensions to}
5593 \c{ELF} object files can contain more information about a global symbol
5594 than just its address: they can contain the \I{symbol sizes,
5595 specifying}\I{size, of symbols}size of the symbol and its \I{symbol
5596 types, specifying}\I{type, of symbols}type as well. These are not
5597 merely debugger conveniences, but are actually necessary when the
5598 program being written is a \i{shared library}. NASM therefore
5599 supports some extensions to the \c{GLOBAL} directive, allowing you
5600 to specify these features.
5602 You can specify whether a global variable is a function or a data
5603 object by suffixing the name with a colon and the word
5604 \i\c{function} or \i\c{data}. (\i\c{object} is a synonym for
5605 \c{data}.) For example:
5607 \c global hashlookup:function, hashtable:data
5609 exports the global symbol \c{hashlookup} as a function and
5610 \c{hashtable} as a data object.
5612 Optionally, you can control the ELF visibility of the symbol. Just
5613 add one of the visibility keywords: \i\c{default}, \i\c{internal},
5614 \i\c{hidden}, or \i\c{protected}. The default is \i\c{default} of
5615 course. For example, to make \c{hashlookup} hidden:
5617 \c global hashlookup:function hidden
5619 You can also specify the size of the data associated with the
5620 symbol, as a numeric expression (which may involve labels, and even
5621 forward references) after the type specifier. Like this:
5623 \c global hashtable:data (hashtable.end - hashtable)
5626 \c db this,that,theother ; some data here
5629 This makes NASM automatically calculate the length of the table and
5630 place that information into the \c{ELF} symbol table.
5632 Declaring the type and size of global symbols is necessary when
5633 writing shared library code. For more information, see
5637 \S{elfcomm} \c{elf} Extensions to the \c{COMMON} Directive
5638 \I{COMMON, elf extensions to}
5640 \c{ELF} also allows you to specify alignment requirements \I{common
5641 variables, alignment in elf}\I{alignment, of elf common variables}on
5642 common variables. This is done by putting a number (which must be a
5643 power of two) after the name and size of the common variable,
5644 separated (as usual) by a colon. For example, an array of
5645 doublewords would benefit from 4-byte alignment:
5647 \c common dwordarray 128:4
5649 This declares the total size of the array to be 128 bytes, and
5650 requires that it be aligned on a 4-byte boundary.
5653 \S{elf16} 16-bit code and ELF
5654 \I{ELF, 16-bit code and}
5656 The \c{ELF32} specification doesn't provide relocations for 8- and
5657 16-bit values, but the GNU \c{ld} linker adds these as an extension.
5658 NASM can generate GNU-compatible relocations, to allow 16-bit code to
5659 be linked as ELF using GNU \c{ld}. If NASM is used with the
5660 \c{-w+gnu-elf-extensions} option, a warning is issued when one of
5661 these relocations is generated.
5663 \S{elfdbg} Debug formats and ELF
5664 \I{ELF, Debug formats and}
5666 \c{ELF32} and \c{ELF64} provide debug information in \c{STABS} and \c{DWARF} formats.
5667 Line number information is generated for all executable sections, but please
5668 note that only the ".text" section is executable by default.
5670 \H{aoutfmt} \i\c{aout}: Linux \I{a.out, Linux version}\I{linux, a.out}\c{a.out} Object Files
5672 The \c{aout} format generates \c{a.out} object files, in the form used
5673 by early Linux systems (current Linux systems use ELF, see
5674 \k{elffmt}.) These differ from other \c{a.out} object files in that
5675 the magic number in the first four bytes of the file is
5676 different; also, some implementations of \c{a.out}, for example
5677 NetBSD's, support position-independent code, which Linux's
5678 implementation does not.
5680 \c{a.out} provides a default output file-name extension of \c{.o}.
5682 \c{a.out} is a very simple object format. It supports no special
5683 directives, no special symbols, no use of \c{SEG} or \c{WRT}, and no
5684 extensions to any standard directives. It supports only the three
5685 \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}.
5688 \H{aoutfmt} \i\c{aoutb}: \i{NetBSD}/\i{FreeBSD}/\i{OpenBSD}
5689 \I{a.out, BSD version}\c{a.out} Object Files
5691 The \c{aoutb} format generates \c{a.out} object files, in the form
5692 used by the various free \c{BSD Unix} clones, \c{NetBSD}, \c{FreeBSD}
5693 and \c{OpenBSD}. For simple object files, this object format is exactly
5694 the same as \c{aout} except for the magic number in the first four bytes
5695 of the file. However, the \c{aoutb} format supports
5696 \I{PIC}\i{position-independent code} in the same way as the \c{elf}
5697 format, so you can use it to write \c{BSD} \i{shared libraries}.
5699 \c{aoutb} provides a default output file-name extension of \c{.o}.
5701 \c{aoutb} supports no special directives, no special symbols, and
5702 only the three \i{standard section names} \i\c{.text}, \i\c{.data}
5703 and \i\c{.bss}. However, it also supports the same use of \i\c{WRT} as
5704 \c{elf} does, to provide position-independent code relocation types.
5705 See \k{elfwrt} for full documentation of this feature.
5707 \c{aoutb} also supports the same extensions to the \c{GLOBAL}
5708 directive as \c{elf} does: see \k{elfglob} for documentation of
5712 \H{as86fmt} \c{as86}: \i{Minix}/Linux\I{linux, as86} \i\c{as86} Object Files
5714 The Minix/Linux 16-bit assembler \c{as86} has its own non-standard
5715 object file format. Although its companion linker \i\c{ld86} produces
5716 something close to ordinary \c{a.out} binaries as output, the object
5717 file format used to communicate between \c{as86} and \c{ld86} is not
5720 NASM supports this format, just in case it is useful, as \c{as86}.
5721 \c{as86} provides a default output file-name extension of \c{.o}.
5723 \c{as86} is a very simple object format (from the NASM user's point
5724 of view). It supports no special directives, no use of \c{SEG} or \c{WRT},
5725 and no extensions to any standard directives. It supports only the three
5726 \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}. The
5727 only special symbol supported is \c{..start}.
5730 \H{rdffmt} \I{RDOFF}\i\c{rdf}: \i{Relocatable Dynamic Object File
5733 The \c{rdf} output format produces \c{RDOFF} object files. \c{RDOFF}
5734 (Relocatable Dynamic Object File Format) is a home-grown object-file
5735 format, designed alongside NASM itself and reflecting in its file
5736 format the internal structure of the assembler.
5738 \c{RDOFF} is not used by any well-known operating systems. Those
5739 writing their own systems, however, may well wish to use \c{RDOFF}
5740 as their object format, on the grounds that it is designed primarily
5741 for simplicity and contains very little file-header bureaucracy.
5743 The Unix NASM archive, and the DOS archive which includes sources,
5744 both contain an \I{rdoff subdirectory}\c{rdoff} subdirectory holding
5745 a set of RDOFF utilities: an RDF linker, an \c{RDF} static-library
5746 manager, an RDF file dump utility, and a program which will load and
5747 execute an RDF executable under Linux.
5749 \c{rdf} supports only the \i{standard section names} \i\c{.text},
5750 \i\c{.data} and \i\c{.bss}.
5753 \S{rdflib} Requiring a Library: The \i\c{LIBRARY} Directive
5755 \c{RDOFF} contains a mechanism for an object file to demand a given
5756 library to be linked to the module, either at load time or run time.
5757 This is done by the \c{LIBRARY} directive, which takes one argument
5758 which is the name of the module:
5760 \c library mylib.rdl
5763 \S{rdfmod} Specifying a Module Name: The \i\c{MODULE} Directive
5765 Special \c{RDOFF} header record is used to store the name of the module.
5766 It can be used, for example, by run-time loader to perform dynamic
5767 linking. \c{MODULE} directive takes one argument which is the name
5772 Note that when you statically link modules and tell linker to strip
5773 the symbols from output file, all module names will be stripped too.
5774 To avoid it, you should start module names with \I{$, prefix}\c{$}, like:
5776 \c module $kernel.core
5779 \S{rdfglob} \c{rdf} Extensions to the \c{GLOBAL} Directive\I{GLOBAL,
5782 \c{RDOFF} global symbols can contain additional information needed by
5783 the static linker. You can mark a global symbol as exported, thus
5784 telling the linker do not strip it from target executable or library
5785 file. Like in \c{ELF}, you can also specify whether an exported symbol
5786 is a procedure (function) or data object.
5788 Suffixing the name with a colon and the word \i\c{export} you make the
5791 \c global sys_open:export
5793 To specify that exported symbol is a procedure (function), you add the
5794 word \i\c{proc} or \i\c{function} after declaration:
5796 \c global sys_open:export proc
5798 Similarly, to specify exported data object, add the word \i\c{data}
5799 or \i\c{object} to the directive:
5801 \c global kernel_ticks:export data
5804 \S{rdfimpt} \c{rdf} Extensions to the \c{EXTERN} Directive\I{EXTERN,
5807 By default the \c{EXTERN} directive in \c{RDOFF} declares a "pure external"
5808 symbol (i.e. the static linker will complain if such a symbol is not resolved).
5809 To declare an "imported" symbol, which must be resolved later during a dynamic
5810 linking phase, \c{RDOFF} offers an additional \c{import} modifier. As in
5811 \c{GLOBAL}, you can also specify whether an imported symbol is a procedure
5812 (function) or data object. For example:
5815 \c extern _open:import
5816 \c extern _printf:import proc
5817 \c extern _errno:import data
5819 Here the directive \c{LIBRARY} is also included, which gives the dynamic linker
5820 a hint as to where to find requested symbols.
5823 \H{dbgfmt} \i\c{dbg}: Debugging Format
5825 The \c{dbg} output format is not built into NASM in the default
5826 configuration. If you are building your own NASM executable from the
5827 sources, you can define \i\c{OF_DBG} in \c{output/outform.h} or on the
5828 compiler command line, and obtain the \c{dbg} output format.
5830 The \c{dbg} format does not output an object file as such; instead,
5831 it outputs a text file which contains a complete list of all the
5832 transactions between the main body of NASM and the output-format
5833 back end module. It is primarily intended to aid people who want to
5834 write their own output drivers, so that they can get a clearer idea
5835 of the various requests the main program makes of the output driver,
5836 and in what order they happen.
5838 For simple files, one can easily use the \c{dbg} format like this:
5840 \c nasm -f dbg filename.asm
5842 which will generate a diagnostic file called \c{filename.dbg}.
5843 However, this will not work well on files which were designed for a
5844 different object format, because each object format defines its own
5845 macros (usually user-level forms of directives), and those macros
5846 will not be defined in the \c{dbg} format. Therefore it can be
5847 useful to run NASM twice, in order to do the preprocessing with the
5848 native object format selected:
5850 \c nasm -e -f rdf -o rdfprog.i rdfprog.asm
5851 \c nasm -a -f dbg rdfprog.i
5853 This preprocesses \c{rdfprog.asm} into \c{rdfprog.i}, keeping the
5854 \c{rdf} object format selected in order to make sure RDF special
5855 directives are converted into primitive form correctly. Then the
5856 preprocessed source is fed through the \c{dbg} format to generate
5857 the final diagnostic output.
5859 This workaround will still typically not work for programs intended
5860 for \c{obj} format, because the \c{obj} \c{SEGMENT} and \c{GROUP}
5861 directives have side effects of defining the segment and group names
5862 as symbols; \c{dbg} will not do this, so the program will not
5863 assemble. You will have to work around that by defining the symbols
5864 yourself (using \c{EXTERN}, for example) if you really need to get a
5865 \c{dbg} trace of an \c{obj}-specific source file.
5867 \c{dbg} accepts any section name and any directives at all, and logs
5868 them all to its output file.
5871 \C{16bit} Writing 16-bit Code (DOS, Windows 3/3.1)
5873 This chapter attempts to cover some of the common issues encountered
5874 when writing 16-bit code to run under \c{MS-DOS} or \c{Windows 3.x}. It
5875 covers how to link programs to produce \c{.EXE} or \c{.COM} files,
5876 how to write \c{.SYS} device drivers, and how to interface assembly
5877 language code with 16-bit C compilers and with Borland Pascal.
5880 \H{exefiles} Producing \i\c{.EXE} Files
5882 Any large program written under DOS needs to be built as a \c{.EXE}
5883 file: only \c{.EXE} files have the necessary internal structure
5884 required to span more than one 64K segment. \i{Windows} programs,
5885 also, have to be built as \c{.EXE} files, since Windows does not
5886 support the \c{.COM} format.
5888 In general, you generate \c{.EXE} files by using the \c{obj} output
5889 format to produce one or more \i\c{.OBJ} files, and then linking
5890 them together using a linker. However, NASM also supports the direct
5891 generation of simple DOS \c{.EXE} files using the \c{bin} output
5892 format (by using \c{DB} and \c{DW} to construct the \c{.EXE} file
5893 header), and a macro package is supplied to do this. Thanks to
5894 Yann Guidon for contributing the code for this.
5896 NASM may also support \c{.EXE} natively as another output format in
5900 \S{objexe} Using the \c{obj} Format To Generate \c{.EXE} Files
5902 This section describes the usual method of generating \c{.EXE} files
5903 by linking \c{.OBJ} files together.
5905 Most 16-bit programming language packages come with a suitable
5906 linker; if you have none of these, there is a free linker called
5907 \i{VAL}\I{linker, free}, available in \c{LZH} archive format from
5908 \W{ftp://x2ftp.oulu.fi/pub/msdos/programming/lang/}\i\c{x2ftp.oulu.fi}.
5909 An LZH archiver can be found at
5910 \W{ftp://ftp.simtel.net/pub/simtelnet/msdos/arcers}\i\c{ftp.simtel.net}.
5911 There is another `free' linker (though this one doesn't come with
5912 sources) called \i{FREELINK}, available from
5913 \W{http://www.pcorner.com/tpc/old/3-101.html}\i\c{www.pcorner.com}.
5914 A third, \i\c{djlink}, written by DJ Delorie, is available at
5915 \W{http://www.delorie.com/djgpp/16bit/djlink/}\i\c{www.delorie.com}.
5916 A fourth linker, \i\c{ALINK}, written by Anthony A.J. Williams, is
5917 available at \W{http://alink.sourceforge.net}\i\c{alink.sourceforge.net}.
5919 When linking several \c{.OBJ} files into a \c{.EXE} file, you should
5920 ensure that exactly one of them has a start point defined (using the
5921 \I{program entry point}\i\c{..start} special symbol defined by the
5922 \c{obj} format: see \k{dotdotstart}). If no module defines a start
5923 point, the linker will not know what value to give the entry-point
5924 field in the output file header; if more than one defines a start
5925 point, the linker will not know \e{which} value to use.
5927 An example of a NASM source file which can be assembled to a
5928 \c{.OBJ} file and linked on its own to a \c{.EXE} is given here. It
5929 demonstrates the basic principles of defining a stack, initialising
5930 the segment registers, and declaring a start point. This file is
5931 also provided in the \I{test subdirectory}\c{test} subdirectory of
5932 the NASM archives, under the name \c{objexe.asm}.
5943 This initial piece of code sets up \c{DS} to point to the data
5944 segment, and initializes \c{SS} and \c{SP} to point to the top of
5945 the provided stack. Notice that interrupts are implicitly disabled
5946 for one instruction after a move into \c{SS}, precisely for this
5947 situation, so that there's no chance of an interrupt occurring
5948 between the loads of \c{SS} and \c{SP} and not having a stack to
5951 Note also that the special symbol \c{..start} is defined at the
5952 beginning of this code, which means that will be the entry point
5953 into the resulting executable file.
5959 The above is the main program: load \c{DS:DX} with a pointer to the
5960 greeting message (\c{hello} is implicitly relative to the segment
5961 \c{data}, which was loaded into \c{DS} in the setup code, so the
5962 full pointer is valid), and call the DOS print-string function.
5967 This terminates the program using another DOS system call.
5971 \c hello: db 'hello, world', 13, 10, '$'
5973 The data segment contains the string we want to display.
5975 \c segment stack stack
5979 The above code declares a stack segment containing 64 bytes of
5980 uninitialized stack space, and points \c{stacktop} at the top of it.
5981 The directive \c{segment stack stack} defines a segment \e{called}
5982 \c{stack}, and also of \e{type} \c{STACK}. The latter is not
5983 necessary to the correct running of the program, but linkers are
5984 likely to issue warnings or errors if your program has no segment of
5987 The above file, when assembled into a \c{.OBJ} file, will link on
5988 its own to a valid \c{.EXE} file, which when run will print `hello,
5989 world' and then exit.
5992 \S{binexe} Using the \c{bin} Format To Generate \c{.EXE} Files
5994 The \c{.EXE} file format is simple enough that it's possible to
5995 build a \c{.EXE} file by writing a pure-binary program and sticking
5996 a 32-byte header on the front. This header is simple enough that it
5997 can be generated using \c{DB} and \c{DW} commands by NASM itself, so
5998 that you can use the \c{bin} output format to directly generate
6001 Included in the NASM archives, in the \I{misc subdirectory}\c{misc}
6002 subdirectory, is a file \i\c{exebin.mac} of macros. It defines three
6003 macros: \i\c{EXE_begin}, \i\c{EXE_stack} and \i\c{EXE_end}.
6005 To produce a \c{.EXE} file using this method, you should start by
6006 using \c{%include} to load the \c{exebin.mac} macro package into
6007 your source file. You should then issue the \c{EXE_begin} macro call
6008 (which takes no arguments) to generate the file header data. Then
6009 write code as normal for the \c{bin} format - you can use all three
6010 standard sections \c{.text}, \c{.data} and \c{.bss}. At the end of
6011 the file you should call the \c{EXE_end} macro (again, no arguments),
6012 which defines some symbols to mark section sizes, and these symbols
6013 are referred to in the header code generated by \c{EXE_begin}.
6015 In this model, the code you end up writing starts at \c{0x100}, just
6016 like a \c{.COM} file - in fact, if you strip off the 32-byte header
6017 from the resulting \c{.EXE} file, you will have a valid \c{.COM}
6018 program. All the segment bases are the same, so you are limited to a
6019 64K program, again just like a \c{.COM} file. Note that an \c{ORG}
6020 directive is issued by the \c{EXE_begin} macro, so you should not
6021 explicitly issue one of your own.
6023 You can't directly refer to your segment base value, unfortunately,
6024 since this would require a relocation in the header, and things
6025 would get a lot more complicated. So you should get your segment
6026 base by copying it out of \c{CS} instead.
6028 On entry to your \c{.EXE} file, \c{SS:SP} are already set up to
6029 point to the top of a 2Kb stack. You can adjust the default stack
6030 size of 2Kb by calling the \c{EXE_stack} macro. For example, to
6031 change the stack size of your program to 64 bytes, you would call
6034 A sample program which generates a \c{.EXE} file in this way is
6035 given in the \c{test} subdirectory of the NASM archive, as
6039 \H{comfiles} Producing \i\c{.COM} Files
6041 While large DOS programs must be written as \c{.EXE} files, small
6042 ones are often better written as \c{.COM} files. \c{.COM} files are
6043 pure binary, and therefore most easily produced using the \c{bin}
6047 \S{combinfmt} Using the \c{bin} Format To Generate \c{.COM} Files
6049 \c{.COM} files expect to be loaded at offset \c{100h} into their
6050 segment (though the segment may change). Execution then begins at
6051 \I\c{ORG}\c{100h}, i.e. right at the start of the program. So to
6052 write a \c{.COM} program, you would create a source file looking
6060 \c ; put your code here
6064 \c ; put data items here
6068 \c ; put uninitialized data here
6070 The \c{bin} format puts the \c{.text} section first in the file, so
6071 you can declare data or BSS items before beginning to write code if
6072 you want to and the code will still end up at the front of the file
6075 The BSS (uninitialized data) section does not take up space in the
6076 \c{.COM} file itself: instead, addresses of BSS items are resolved
6077 to point at space beyond the end of the file, on the grounds that
6078 this will be free memory when the program is run. Therefore you
6079 should not rely on your BSS being initialized to all zeros when you
6082 To assemble the above program, you should use a command line like
6084 \c nasm myprog.asm -fbin -o myprog.com
6086 The \c{bin} format would produce a file called \c{myprog} if no
6087 explicit output file name were specified, so you have to override it
6088 and give the desired file name.
6091 \S{comobjfmt} Using the \c{obj} Format To Generate \c{.COM} Files
6093 If you are writing a \c{.COM} program as more than one module, you
6094 may wish to assemble several \c{.OBJ} files and link them together
6095 into a \c{.COM} program. You can do this, provided you have a linker
6096 capable of outputting \c{.COM} files directly (\i{TLINK} does this),
6097 or alternatively a converter program such as \i\c{EXE2BIN} to
6098 transform the \c{.EXE} file output from the linker into a \c{.COM}
6101 If you do this, you need to take care of several things:
6103 \b The first object file containing code should start its code
6104 segment with a line like \c{RESB 100h}. This is to ensure that the
6105 code begins at offset \c{100h} relative to the beginning of the code
6106 segment, so that the linker or converter program does not have to
6107 adjust address references within the file when generating the
6108 \c{.COM} file. Other assemblers use an \i\c{ORG} directive for this
6109 purpose, but \c{ORG} in NASM is a format-specific directive to the
6110 \c{bin} output format, and does not mean the same thing as it does
6111 in MASM-compatible assemblers.
6113 \b You don't need to define a stack segment.
6115 \b All your segments should be in the same group, so that every time
6116 your code or data references a symbol offset, all offsets are
6117 relative to the same segment base. This is because, when a \c{.COM}
6118 file is loaded, all the segment registers contain the same value.
6121 \H{sysfiles} Producing \i\c{.SYS} Files
6123 \i{MS-DOS device drivers} - \c{.SYS} files - are pure binary files,
6124 similar to \c{.COM} files, except that they start at origin zero
6125 rather than \c{100h}. Therefore, if you are writing a device driver
6126 using the \c{bin} format, you do not need the \c{ORG} directive,
6127 since the default origin for \c{bin} is zero. Similarly, if you are
6128 using \c{obj}, you do not need the \c{RESB 100h} at the start of
6131 \c{.SYS} files start with a header structure, containing pointers to
6132 the various routines inside the driver which do the work. This
6133 structure should be defined at the start of the code segment, even
6134 though it is not actually code.
6136 For more information on the format of \c{.SYS} files, and the data
6137 which has to go in the header structure, a list of books is given in
6138 the Frequently Asked Questions list for the newsgroup
6139 \W{news:comp.os.msdos.programmer}\i\c{comp.os.msdos.programmer}.
6142 \H{16c} Interfacing to 16-bit C Programs
6144 This section covers the basics of writing assembly routines that
6145 call, or are called from, C programs. To do this, you would
6146 typically write an assembly module as a \c{.OBJ} file, and link it
6147 with your C modules to produce a \i{mixed-language program}.
6150 \S{16cunder} External Symbol Names
6152 \I{C symbol names}\I{underscore, in C symbols}C compilers have the
6153 convention that the names of all global symbols (functions or data)
6154 they define are formed by prefixing an underscore to the name as it
6155 appears in the C program. So, for example, the function a C
6156 programmer thinks of as \c{printf} appears to an assembly language
6157 programmer as \c{_printf}. This means that in your assembly
6158 programs, you can define symbols without a leading underscore, and
6159 not have to worry about name clashes with C symbols.
6161 If you find the underscores inconvenient, you can define macros to
6162 replace the \c{GLOBAL} and \c{EXTERN} directives as follows:
6178 (These forms of the macros only take one argument at a time; a
6179 \c{%rep} construct could solve this.)
6181 If you then declare an external like this:
6185 then the macro will expand it as
6188 \c %define printf _printf
6190 Thereafter, you can reference \c{printf} as if it was a symbol, and
6191 the preprocessor will put the leading underscore on where necessary.
6193 The \c{cglobal} macro works similarly. You must use \c{cglobal}
6194 before defining the symbol in question, but you would have had to do
6195 that anyway if you used \c{GLOBAL}.
6197 Also see \k{opt-pfix}.
6199 \S{16cmodels} \i{Memory Models}
6201 NASM contains no mechanism to support the various C memory models
6202 directly; you have to keep track yourself of which one you are
6203 writing for. This means you have to keep track of the following
6206 \b In models using a single code segment (tiny, small and compact),
6207 functions are near. This means that function pointers, when stored
6208 in data segments or pushed on the stack as function arguments, are
6209 16 bits long and contain only an offset field (the \c{CS} register
6210 never changes its value, and always gives the segment part of the
6211 full function address), and that functions are called using ordinary
6212 near \c{CALL} instructions and return using \c{RETN} (which, in
6213 NASM, is synonymous with \c{RET} anyway). This means both that you
6214 should write your own routines to return with \c{RETN}, and that you
6215 should call external C routines with near \c{CALL} instructions.
6217 \b In models using more than one code segment (medium, large and
6218 huge), functions are far. This means that function pointers are 32
6219 bits long (consisting of a 16-bit offset followed by a 16-bit
6220 segment), and that functions are called using \c{CALL FAR} (or
6221 \c{CALL seg:offset}) and return using \c{RETF}. Again, you should
6222 therefore write your own routines to return with \c{RETF} and use
6223 \c{CALL FAR} to call external routines.
6225 \b In models using a single data segment (tiny, small and medium),
6226 data pointers are 16 bits long, containing only an offset field (the
6227 \c{DS} register doesn't change its value, and always gives the
6228 segment part of the full data item address).
6230 \b In models using more than one data segment (compact, large and
6231 huge), data pointers are 32 bits long, consisting of a 16-bit offset
6232 followed by a 16-bit segment. You should still be careful not to
6233 modify \c{DS} in your routines without restoring it afterwards, but
6234 \c{ES} is free for you to use to access the contents of 32-bit data
6235 pointers you are passed.
6237 \b The huge memory model allows single data items to exceed 64K in
6238 size. In all other memory models, you can access the whole of a data
6239 item just by doing arithmetic on the offset field of the pointer you
6240 are given, whether a segment field is present or not; in huge model,
6241 you have to be more careful of your pointer arithmetic.
6243 \b In most memory models, there is a \e{default} data segment, whose
6244 segment address is kept in \c{DS} throughout the program. This data
6245 segment is typically the same segment as the stack, kept in \c{SS},
6246 so that functions' local variables (which are stored on the stack)
6247 and global data items can both be accessed easily without changing
6248 \c{DS}. Particularly large data items are typically stored in other
6249 segments. However, some memory models (though not the standard
6250 ones, usually) allow the assumption that \c{SS} and \c{DS} hold the
6251 same value to be removed. Be careful about functions' local
6252 variables in this latter case.
6254 In models with a single code segment, the segment is called
6255 \i\c{_TEXT}, so your code segment must also go by this name in order
6256 to be linked into the same place as the main code segment. In models
6257 with a single data segment, or with a default data segment, it is
6261 \S{16cfunc} Function Definitions and Function Calls
6263 \I{functions, C calling convention}The \i{C calling convention} in
6264 16-bit programs is as follows. In the following description, the
6265 words \e{caller} and \e{callee} are used to denote the function
6266 doing the calling and the function which gets called.
6268 \b The caller pushes the function's parameters on the stack, one
6269 after another, in reverse order (right to left, so that the first
6270 argument specified to the function is pushed last).
6272 \b The caller then executes a \c{CALL} instruction to pass control
6273 to the callee. This \c{CALL} is either near or far depending on the
6276 \b The callee receives control, and typically (although this is not
6277 actually necessary, in functions which do not need to access their
6278 parameters) starts by saving the value of \c{SP} in \c{BP} so as to
6279 be able to use \c{BP} as a base pointer to find its parameters on
6280 the stack. However, the caller was probably doing this too, so part
6281 of the calling convention states that \c{BP} must be preserved by
6282 any C function. Hence the callee, if it is going to set up \c{BP} as
6283 a \i\e{frame pointer}, must push the previous value first.
6285 \b The callee may then access its parameters relative to \c{BP}.
6286 The word at \c{[BP]} holds the previous value of \c{BP} as it was
6287 pushed; the next word, at \c{[BP+2]}, holds the offset part of the
6288 return address, pushed implicitly by \c{CALL}. In a small-model
6289 (near) function, the parameters start after that, at \c{[BP+4]}; in
6290 a large-model (far) function, the segment part of the return address
6291 lives at \c{[BP+4]}, and the parameters begin at \c{[BP+6]}. The
6292 leftmost parameter of the function, since it was pushed last, is
6293 accessible at this offset from \c{BP}; the others follow, at
6294 successively greater offsets. Thus, in a function such as \c{printf}
6295 which takes a variable number of parameters, the pushing of the
6296 parameters in reverse order means that the function knows where to
6297 find its first parameter, which tells it the number and type of the
6300 \b The callee may also wish to decrease \c{SP} further, so as to
6301 allocate space on the stack for local variables, which will then be
6302 accessible at negative offsets from \c{BP}.
6304 \b The callee, if it wishes to return a value to the caller, should
6305 leave the value in \c{AL}, \c{AX} or \c{DX:AX} depending on the size
6306 of the value. Floating-point results are sometimes (depending on the
6307 compiler) returned in \c{ST0}.
6309 \b Once the callee has finished processing, it restores \c{SP} from
6310 \c{BP} if it had allocated local stack space, then pops the previous
6311 value of \c{BP}, and returns via \c{RETN} or \c{RETF} depending on
6314 \b When the caller regains control from the callee, the function
6315 parameters are still on the stack, so it typically adds an immediate
6316 constant to \c{SP} to remove them (instead of executing a number of
6317 slow \c{POP} instructions). Thus, if a function is accidentally
6318 called with the wrong number of parameters due to a prototype
6319 mismatch, the stack will still be returned to a sensible state since
6320 the caller, which \e{knows} how many parameters it pushed, does the
6323 It is instructive to compare this calling convention with that for
6324 Pascal programs (described in \k{16bpfunc}). Pascal has a simpler
6325 convention, since no functions have variable numbers of parameters.
6326 Therefore the callee knows how many parameters it should have been
6327 passed, and is able to deallocate them from the stack itself by
6328 passing an immediate argument to the \c{RET} or \c{RETF}
6329 instruction, so the caller does not have to do it. Also, the
6330 parameters are pushed in left-to-right order, not right-to-left,
6331 which means that a compiler can give better guarantees about
6332 sequence points without performance suffering.
6334 Thus, you would define a function in C style in the following way.
6335 The following example is for small model:
6342 \c sub sp,0x40 ; 64 bytes of local stack space
6343 \c mov bx,[bp+4] ; first parameter to function
6347 \c mov sp,bp ; undo "sub sp,0x40" above
6351 For a large-model function, you would replace \c{RET} by \c{RETF},
6352 and look for the first parameter at \c{[BP+6]} instead of
6353 \c{[BP+4]}. Of course, if one of the parameters is a pointer, then
6354 the offsets of \e{subsequent} parameters will change depending on
6355 the memory model as well: far pointers take up four bytes on the
6356 stack when passed as a parameter, whereas near pointers take up two.
6358 At the other end of the process, to call a C function from your
6359 assembly code, you would do something like this:
6363 \c ; and then, further down...
6365 \c push word [myint] ; one of my integer variables
6366 \c push word mystring ; pointer into my data segment
6368 \c add sp,byte 4 ; `byte' saves space
6370 \c ; then those data items...
6375 \c mystring db 'This number -> %d <- should be 1234',10,0
6377 This piece of code is the small-model assembly equivalent of the C
6380 \c int myint = 1234;
6381 \c printf("This number -> %d <- should be 1234\n", myint);
6383 In large model, the function-call code might look more like this. In
6384 this example, it is assumed that \c{DS} already holds the segment
6385 base of the segment \c{_DATA}. If not, you would have to initialize
6388 \c push word [myint]
6389 \c push word seg mystring ; Now push the segment, and...
6390 \c push word mystring ; ... offset of "mystring"
6394 The integer value still takes up one word on the stack, since large
6395 model does not affect the size of the \c{int} data type. The first
6396 argument (pushed last) to \c{printf}, however, is a data pointer,
6397 and therefore has to contain a segment and offset part. The segment
6398 should be stored second in memory, and therefore must be pushed
6399 first. (Of course, \c{PUSH DS} would have been a shorter instruction
6400 than \c{PUSH WORD SEG mystring}, if \c{DS} was set up as the above
6401 example assumed.) Then the actual call becomes a far call, since
6402 functions expect far calls in large model; and \c{SP} has to be
6403 increased by 6 rather than 4 afterwards to make up for the extra
6407 \S{16cdata} Accessing Data Items
6409 To get at the contents of C variables, or to declare variables which
6410 C can access, you need only declare the names as \c{GLOBAL} or
6411 \c{EXTERN}. (Again, the names require leading underscores, as stated
6412 in \k{16cunder}.) Thus, a C variable declared as \c{int i} can be
6413 accessed from assembler as
6419 And to declare your own integer variable which C programs can access
6420 as \c{extern int j}, you do this (making sure you are assembling in
6421 the \c{_DATA} segment, if necessary):
6427 To access a C array, you need to know the size of the components of
6428 the array. For example, \c{int} variables are two bytes long, so if
6429 a C program declares an array as \c{int a[10]}, you can access
6430 \c{a[3]} by coding \c{mov ax,[_a+6]}. (The byte offset 6 is obtained
6431 by multiplying the desired array index, 3, by the size of the array
6432 element, 2.) The sizes of the C base types in 16-bit compilers are:
6433 1 for \c{char}, 2 for \c{short} and \c{int}, 4 for \c{long} and
6434 \c{float}, and 8 for \c{double}.
6436 To access a C \i{data structure}, you need to know the offset from
6437 the base of the structure to the field you are interested in. You
6438 can either do this by converting the C structure definition into a
6439 NASM structure definition (using \i\c{STRUC}), or by calculating the
6440 one offset and using just that.
6442 To do either of these, you should read your C compiler's manual to
6443 find out how it organizes data structures. NASM gives no special
6444 alignment to structure members in its own \c{STRUC} macro, so you
6445 have to specify alignment yourself if the C compiler generates it.
6446 Typically, you might find that a structure like
6453 might be four bytes long rather than three, since the \c{int} field
6454 would be aligned to a two-byte boundary. However, this sort of
6455 feature tends to be a configurable option in the C compiler, either
6456 using command-line options or \c{#pragma} lines, so you have to find
6457 out how your own compiler does it.
6460 \S{16cmacro} \i\c{c16.mac}: Helper Macros for the 16-bit C Interface
6462 Included in the NASM archives, in the \I{misc subdirectory}\c{misc}
6463 directory, is a file \c{c16.mac} of macros. It defines three macros:
6464 \i\c{proc}, \i\c{arg} and \i\c{endproc}. These are intended to be
6465 used for C-style procedure definitions, and they automate a lot of
6466 the work involved in keeping track of the calling convention.
6468 (An alternative, TASM compatible form of \c{arg} is also now built
6469 into NASM's preprocessor. See \k{stackrel} for details.)
6471 An example of an assembly function using the macro set is given
6478 \c mov ax,[bp + %$i]
6479 \c mov bx,[bp + %$j]
6484 This defines \c{_nearproc} to be a procedure taking two arguments,
6485 the first (\c{i}) an integer and the second (\c{j}) a pointer to an
6486 integer. It returns \c{i + *j}.
6488 Note that the \c{arg} macro has an \c{EQU} as the first line of its
6489 expansion, and since the label before the macro call gets prepended
6490 to the first line of the expanded macro, the \c{EQU} works, defining
6491 \c{%$i} to be an offset from \c{BP}. A context-local variable is
6492 used, local to the context pushed by the \c{proc} macro and popped
6493 by the \c{endproc} macro, so that the same argument name can be used
6494 in later procedures. Of course, you don't \e{have} to do that.
6496 The macro set produces code for near functions (tiny, small and
6497 compact-model code) by default. You can have it generate far
6498 functions (medium, large and huge-model code) by means of coding
6499 \I\c{FARCODE}\c{%define FARCODE}. This changes the kind of return
6500 instruction generated by \c{endproc}, and also changes the starting
6501 point for the argument offsets. The macro set contains no intrinsic
6502 dependency on whether data pointers are far or not.
6504 \c{arg} can take an optional parameter, giving the size of the
6505 argument. If no size is given, 2 is assumed, since it is likely that
6506 many function parameters will be of type \c{int}.
6508 The large-model equivalent of the above function would look like this:
6516 \c mov ax,[bp + %$i]
6517 \c mov bx,[bp + %$j]
6518 \c mov es,[bp + %$j + 2]
6523 This makes use of the argument to the \c{arg} macro to define a
6524 parameter of size 4, because \c{j} is now a far pointer. When we
6525 load from \c{j}, we must load a segment and an offset.
6528 \H{16bp} Interfacing to \i{Borland Pascal} Programs
6530 Interfacing to Borland Pascal programs is similar in concept to
6531 interfacing to 16-bit C programs. The differences are:
6533 \b The leading underscore required for interfacing to C programs is
6534 not required for Pascal.
6536 \b The memory model is always large: functions are far, data
6537 pointers are far, and no data item can be more than 64K long.
6538 (Actually, some functions are near, but only those functions that
6539 are local to a Pascal unit and never called from outside it. All
6540 assembly functions that Pascal calls, and all Pascal functions that
6541 assembly routines are able to call, are far.) However, all static
6542 data declared in a Pascal program goes into the default data
6543 segment, which is the one whose segment address will be in \c{DS}
6544 when control is passed to your assembly code. The only things that
6545 do not live in the default data segment are local variables (they
6546 live in the stack segment) and dynamically allocated variables. All
6547 data \e{pointers}, however, are far.
6549 \b The function calling convention is different - described below.
6551 \b Some data types, such as strings, are stored differently.
6553 \b There are restrictions on the segment names you are allowed to
6554 use - Borland Pascal will ignore code or data declared in a segment
6555 it doesn't like the name of. The restrictions are described below.
6558 \S{16bpfunc} The Pascal Calling Convention
6560 \I{functions, Pascal calling convention}\I{Pascal calling
6561 convention}The 16-bit Pascal calling convention is as follows. In
6562 the following description, the words \e{caller} and \e{callee} are
6563 used to denote the function doing the calling and the function which
6566 \b The caller pushes the function's parameters on the stack, one
6567 after another, in normal order (left to right, so that the first
6568 argument specified to the function is pushed first).
6570 \b The caller then executes a far \c{CALL} instruction to pass
6571 control to the callee.
6573 \b The callee receives control, and typically (although this is not
6574 actually necessary, in functions which do not need to access their
6575 parameters) starts by saving the value of \c{SP} in \c{BP} so as to
6576 be able to use \c{BP} as a base pointer to find its parameters on
6577 the stack. However, the caller was probably doing this too, so part
6578 of the calling convention states that \c{BP} must be preserved by
6579 any function. Hence the callee, if it is going to set up \c{BP} as a
6580 \i{frame pointer}, must push the previous value first.
6582 \b The callee may then access its parameters relative to \c{BP}.
6583 The word at \c{[BP]} holds the previous value of \c{BP} as it was
6584 pushed. The next word, at \c{[BP+2]}, holds the offset part of the
6585 return address, and the next one at \c{[BP+4]} the segment part. The
6586 parameters begin at \c{[BP+6]}. The rightmost parameter of the
6587 function, since it was pushed last, is accessible at this offset
6588 from \c{BP}; the others follow, at successively greater offsets.
6590 \b The callee may also wish to decrease \c{SP} further, so as to
6591 allocate space on the stack for local variables, which will then be
6592 accessible at negative offsets from \c{BP}.
6594 \b The callee, if it wishes to return a value to the caller, should
6595 leave the value in \c{AL}, \c{AX} or \c{DX:AX} depending on the size
6596 of the value. Floating-point results are returned in \c{ST0}.
6597 Results of type \c{Real} (Borland's own custom floating-point data
6598 type, not handled directly by the FPU) are returned in \c{DX:BX:AX}.
6599 To return a result of type \c{String}, the caller pushes a pointer
6600 to a temporary string before pushing the parameters, and the callee
6601 places the returned string value at that location. The pointer is
6602 not a parameter, and should not be removed from the stack by the
6603 \c{RETF} instruction.
6605 \b Once the callee has finished processing, it restores \c{SP} from
6606 \c{BP} if it had allocated local stack space, then pops the previous
6607 value of \c{BP}, and returns via \c{RETF}. It uses the form of
6608 \c{RETF} with an immediate parameter, giving the number of bytes
6609 taken up by the parameters on the stack. This causes the parameters
6610 to be removed from the stack as a side effect of the return
6613 \b When the caller regains control from the callee, the function
6614 parameters have already been removed from the stack, so it needs to
6617 Thus, you would define a function in Pascal style, taking two
6618 \c{Integer}-type parameters, in the following way:
6624 \c sub sp,0x40 ; 64 bytes of local stack space
6625 \c mov bx,[bp+8] ; first parameter to function
6626 \c mov bx,[bp+6] ; second parameter to function
6630 \c mov sp,bp ; undo "sub sp,0x40" above
6632 \c retf 4 ; total size of params is 4
6634 At the other end of the process, to call a Pascal function from your
6635 assembly code, you would do something like this:
6639 \c ; and then, further down...
6641 \c push word seg mystring ; Now push the segment, and...
6642 \c push word mystring ; ... offset of "mystring"
6643 \c push word [myint] ; one of my variables
6644 \c call far SomeFunc
6646 This is equivalent to the Pascal code
6648 \c procedure SomeFunc(String: PChar; Int: Integer);
6649 \c SomeFunc(@mystring, myint);
6652 \S{16bpseg} Borland Pascal \I{segment names, Borland Pascal}Segment
6655 Since Borland Pascal's internal unit file format is completely
6656 different from \c{OBJ}, it only makes a very sketchy job of actually
6657 reading and understanding the various information contained in a
6658 real \c{OBJ} file when it links that in. Therefore an object file
6659 intended to be linked to a Pascal program must obey a number of
6662 \b Procedures and functions must be in a segment whose name is
6663 either \c{CODE}, \c{CSEG}, or something ending in \c{_TEXT}.
6665 \b initialized data must be in a segment whose name is either
6666 \c{CONST} or something ending in \c{_DATA}.
6668 \b Uninitialized data must be in a segment whose name is either
6669 \c{DATA}, \c{DSEG}, or something ending in \c{_BSS}.
6671 \b Any other segments in the object file are completely ignored.
6672 \c{GROUP} directives and segment attributes are also ignored.
6675 \S{16bpmacro} Using \i\c{c16.mac} With Pascal Programs
6677 The \c{c16.mac} macro package, described in \k{16cmacro}, can also
6678 be used to simplify writing functions to be called from Pascal
6679 programs, if you code \I\c{PASCAL}\c{%define PASCAL}. This
6680 definition ensures that functions are far (it implies
6681 \i\c{FARCODE}), and also causes procedure return instructions to be
6682 generated with an operand.
6684 Defining \c{PASCAL} does not change the code which calculates the
6685 argument offsets; you must declare your function's arguments in
6686 reverse order. For example:
6694 \c mov ax,[bp + %$i]
6695 \c mov bx,[bp + %$j]
6696 \c mov es,[bp + %$j + 2]
6701 This defines the same routine, conceptually, as the example in
6702 \k{16cmacro}: it defines a function taking two arguments, an integer
6703 and a pointer to an integer, which returns the sum of the integer
6704 and the contents of the pointer. The only difference between this
6705 code and the large-model C version is that \c{PASCAL} is defined
6706 instead of \c{FARCODE}, and that the arguments are declared in
6710 \C{32bit} Writing 32-bit Code (Unix, Win32, DJGPP)
6712 This chapter attempts to cover some of the common issues involved
6713 when writing 32-bit code, to run under \i{Win32} or Unix, or to be
6714 linked with C code generated by a Unix-style C compiler such as
6715 \i{DJGPP}. It covers how to write assembly code to interface with
6716 32-bit C routines, and how to write position-independent code for
6719 Almost all 32-bit code, and in particular all code running under
6720 \c{Win32}, \c{DJGPP} or any of the PC Unix variants, runs in \I{flat
6721 memory model}\e{flat} memory model. This means that the segment registers
6722 and paging have already been set up to give you the same 32-bit 4Gb
6723 address space no matter what segment you work relative to, and that
6724 you should ignore all segment registers completely. When writing
6725 flat-model application code, you never need to use a segment
6726 override or modify any segment register, and the code-section
6727 addresses you pass to \c{CALL} and \c{JMP} live in the same address
6728 space as the data-section addresses you access your variables by and
6729 the stack-section addresses you access local variables and procedure
6730 parameters by. Every address is 32 bits long and contains only an
6734 \H{32c} Interfacing to 32-bit C Programs
6736 A lot of the discussion in \k{16c}, about interfacing to 16-bit C
6737 programs, still applies when working in 32 bits. The absence of
6738 memory models or segmentation worries simplifies things a lot.
6741 \S{32cunder} External Symbol Names
6743 Most 32-bit C compilers share the convention used by 16-bit
6744 compilers, that the names of all global symbols (functions or data)
6745 they define are formed by prefixing an underscore to the name as it
6746 appears in the C program. However, not all of them do: the \c{ELF}
6747 specification states that C symbols do \e{not} have a leading
6748 underscore on their assembly-language names.
6750 The older Linux \c{a.out} C compiler, all \c{Win32} compilers,
6751 \c{DJGPP}, and \c{NetBSD} and \c{FreeBSD}, all use the leading
6752 underscore; for these compilers, the macros \c{cextern} and
6753 \c{cglobal}, as given in \k{16cunder}, will still work. For \c{ELF},
6754 though, the leading underscore should not be used.
6756 See also \k{opt-pfix}.
6758 \S{32cfunc} Function Definitions and Function Calls
6760 \I{functions, C calling convention}The \i{C calling convention}
6761 in 32-bit programs is as follows. In the following description,
6762 the words \e{caller} and \e{callee} are used to denote
6763 the function doing the calling and the function which gets called.
6765 \b The caller pushes the function's parameters on the stack, one
6766 after another, in reverse order (right to left, so that the first
6767 argument specified to the function is pushed last).
6769 \b The caller then executes a near \c{CALL} instruction to pass
6770 control to the callee.
6772 \b The callee receives control, and typically (although this is not
6773 actually necessary, in functions which do not need to access their
6774 parameters) starts by saving the value of \c{ESP} in \c{EBP} so as
6775 to be able to use \c{EBP} as a base pointer to find its parameters
6776 on the stack. However, the caller was probably doing this too, so
6777 part of the calling convention states that \c{EBP} must be preserved
6778 by any C function. Hence the callee, if it is going to set up
6779 \c{EBP} as a \i{frame pointer}, must push the previous value first.
6781 \b The callee may then access its parameters relative to \c{EBP}.
6782 The doubleword at \c{[EBP]} holds the previous value of \c{EBP} as
6783 it was pushed; the next doubleword, at \c{[EBP+4]}, holds the return
6784 address, pushed implicitly by \c{CALL}. The parameters start after
6785 that, at \c{[EBP+8]}. The leftmost parameter of the function, since
6786 it was pushed last, is accessible at this offset from \c{EBP}; the
6787 others follow, at successively greater offsets. Thus, in a function
6788 such as \c{printf} which takes a variable number of parameters, the
6789 pushing of the parameters in reverse order means that the function
6790 knows where to find its first parameter, which tells it the number
6791 and type of the remaining ones.
6793 \b The callee may also wish to decrease \c{ESP} further, so as to
6794 allocate space on the stack for local variables, which will then be
6795 accessible at negative offsets from \c{EBP}.
6797 \b The callee, if it wishes to return a value to the caller, should
6798 leave the value in \c{AL}, \c{AX} or \c{EAX} depending on the size
6799 of the value. Floating-point results are typically returned in
6802 \b Once the callee has finished processing, it restores \c{ESP} from
6803 \c{EBP} if it had allocated local stack space, then pops the previous
6804 value of \c{EBP}, and returns via \c{RET} (equivalently, \c{RETN}).
6806 \b When the caller regains control from the callee, the function
6807 parameters are still on the stack, so it typically adds an immediate
6808 constant to \c{ESP} to remove them (instead of executing a number of
6809 slow \c{POP} instructions). Thus, if a function is accidentally
6810 called with the wrong number of parameters due to a prototype
6811 mismatch, the stack will still be returned to a sensible state since
6812 the caller, which \e{knows} how many parameters it pushed, does the
6815 There is an alternative calling convention used by Win32 programs
6816 for Windows API calls, and also for functions called \e{by} the
6817 Windows API such as window procedures: they follow what Microsoft
6818 calls the \c{__stdcall} convention. This is slightly closer to the
6819 Pascal convention, in that the callee clears the stack by passing a
6820 parameter to the \c{RET} instruction. However, the parameters are
6821 still pushed in right-to-left order.
6823 Thus, you would define a function in C style in the following way:
6830 \c sub esp,0x40 ; 64 bytes of local stack space
6831 \c mov ebx,[ebp+8] ; first parameter to function
6835 \c leave ; mov esp,ebp / pop ebp
6838 At the other end of the process, to call a C function from your
6839 assembly code, you would do something like this:
6843 \c ; and then, further down...
6845 \c push dword [myint] ; one of my integer variables
6846 \c push dword mystring ; pointer into my data segment
6848 \c add esp,byte 8 ; `byte' saves space
6850 \c ; then those data items...
6855 \c mystring db 'This number -> %d <- should be 1234',10,0
6857 This piece of code is the assembly equivalent of the C code
6859 \c int myint = 1234;
6860 \c printf("This number -> %d <- should be 1234\n", myint);
6863 \S{32cdata} Accessing Data Items
6865 To get at the contents of C variables, or to declare variables which
6866 C can access, you need only declare the names as \c{GLOBAL} or
6867 \c{EXTERN}. (Again, the names require leading underscores, as stated
6868 in \k{32cunder}.) Thus, a C variable declared as \c{int i} can be
6869 accessed from assembler as
6874 And to declare your own integer variable which C programs can access
6875 as \c{extern int j}, you do this (making sure you are assembling in
6876 the \c{_DATA} segment, if necessary):
6881 To access a C array, you need to know the size of the components of
6882 the array. For example, \c{int} variables are four bytes long, so if
6883 a C program declares an array as \c{int a[10]}, you can access
6884 \c{a[3]} by coding \c{mov ax,[_a+12]}. (The byte offset 12 is obtained
6885 by multiplying the desired array index, 3, by the size of the array
6886 element, 4.) The sizes of the C base types in 32-bit compilers are:
6887 1 for \c{char}, 2 for \c{short}, 4 for \c{int}, \c{long} and
6888 \c{float}, and 8 for \c{double}. Pointers, being 32-bit addresses,
6889 are also 4 bytes long.
6891 To access a C \i{data structure}, you need to know the offset from
6892 the base of the structure to the field you are interested in. You
6893 can either do this by converting the C structure definition into a
6894 NASM structure definition (using \c{STRUC}), or by calculating the
6895 one offset and using just that.
6897 To do either of these, you should read your C compiler's manual to
6898 find out how it organizes data structures. NASM gives no special
6899 alignment to structure members in its own \i\c{STRUC} macro, so you
6900 have to specify alignment yourself if the C compiler generates it.
6901 Typically, you might find that a structure like
6908 might be eight bytes long rather than five, since the \c{int} field
6909 would be aligned to a four-byte boundary. However, this sort of
6910 feature is sometimes a configurable option in the C compiler, either
6911 using command-line options or \c{#pragma} lines, so you have to find
6912 out how your own compiler does it.
6915 \S{32cmacro} \i\c{c32.mac}: Helper Macros for the 32-bit C Interface
6917 Included in the NASM archives, in the \I{misc directory}\c{misc}
6918 directory, is a file \c{c32.mac} of macros. It defines three macros:
6919 \i\c{proc}, \i\c{arg} and \i\c{endproc}. These are intended to be
6920 used for C-style procedure definitions, and they automate a lot of
6921 the work involved in keeping track of the calling convention.
6923 An example of an assembly function using the macro set is given
6930 \c mov eax,[ebp + %$i]
6931 \c mov ebx,[ebp + %$j]
6936 This defines \c{_proc32} to be a procedure taking two arguments, the
6937 first (\c{i}) an integer and the second (\c{j}) a pointer to an
6938 integer. It returns \c{i + *j}.
6940 Note that the \c{arg} macro has an \c{EQU} as the first line of its
6941 expansion, and since the label before the macro call gets prepended
6942 to the first line of the expanded macro, the \c{EQU} works, defining
6943 \c{%$i} to be an offset from \c{BP}. A context-local variable is
6944 used, local to the context pushed by the \c{proc} macro and popped
6945 by the \c{endproc} macro, so that the same argument name can be used
6946 in later procedures. Of course, you don't \e{have} to do that.
6948 \c{arg} can take an optional parameter, giving the size of the
6949 argument. If no size is given, 4 is assumed, since it is likely that
6950 many function parameters will be of type \c{int} or pointers.
6953 \H{picdll} Writing NetBSD/FreeBSD/OpenBSD and Linux/ELF \i{Shared
6956 \c{ELF} replaced the older \c{a.out} object file format under Linux
6957 because it contains support for \i{position-independent code}
6958 (\i{PIC}), which makes writing shared libraries much easier. NASM
6959 supports the \c{ELF} position-independent code features, so you can
6960 write Linux \c{ELF} shared libraries in NASM.
6962 \i{NetBSD}, and its close cousins \i{FreeBSD} and \i{OpenBSD}, take
6963 a different approach by hacking PIC support into the \c{a.out}
6964 format. NASM supports this as the \i\c{aoutb} output format, so you
6965 can write \i{BSD} shared libraries in NASM too.
6967 The operating system loads a PIC shared library by memory-mapping
6968 the library file at an arbitrarily chosen point in the address space
6969 of the running process. The contents of the library's code section
6970 must therefore not depend on where it is loaded in memory.
6972 Therefore, you cannot get at your variables by writing code like
6975 \c mov eax,[myvar] ; WRONG
6977 Instead, the linker provides an area of memory called the
6978 \i\e{global offset table}, or \i{GOT}; the GOT is situated at a
6979 constant distance from your library's code, so if you can find out
6980 where your library is loaded (which is typically done using a
6981 \c{CALL} and \c{POP} combination), you can obtain the address of the
6982 GOT, and you can then load the addresses of your variables out of
6983 linker-generated entries in the GOT.
6985 The \e{data} section of a PIC shared library does not have these
6986 restrictions: since the data section is writable, it has to be
6987 copied into memory anyway rather than just paged in from the library
6988 file, so as long as it's being copied it can be relocated too. So
6989 you can put ordinary types of relocation in the data section without
6990 too much worry (but see \k{picglobal} for a caveat).
6993 \S{picgot} Obtaining the Address of the GOT
6995 Each code module in your shared library should define the GOT as an
6998 \c extern _GLOBAL_OFFSET_TABLE_ ; in ELF
6999 \c extern __GLOBAL_OFFSET_TABLE_ ; in BSD a.out
7001 At the beginning of any function in your shared library which plans
7002 to access your data or BSS sections, you must first calculate the
7003 address of the GOT. This is typically done by writing the function
7012 \c add ebx,_GLOBAL_OFFSET_TABLE_+$$-.get_GOT wrt ..gotpc
7014 \c ; the function body comes here
7021 (For BSD, again, the symbol \c{_GLOBAL_OFFSET_TABLE} requires a
7022 second leading underscore.)
7024 The first two lines of this function are simply the standard C
7025 prologue to set up a stack frame, and the last three lines are
7026 standard C function epilogue. The third line, and the fourth to last
7027 line, save and restore the \c{EBX} register, because PIC shared
7028 libraries use this register to store the address of the GOT.
7030 The interesting bit is the \c{CALL} instruction and the following
7031 two lines. The \c{CALL} and \c{POP} combination obtains the address
7032 of the label \c{.get_GOT}, without having to know in advance where
7033 the program was loaded (since the \c{CALL} instruction is encoded
7034 relative to the current position). The \c{ADD} instruction makes use
7035 of one of the special PIC relocation types: \i{GOTPC relocation}.
7036 With the \i\c{WRT ..gotpc} qualifier specified, the symbol
7037 referenced (here \c{_GLOBAL_OFFSET_TABLE_}, the special symbol
7038 assigned to the GOT) is given as an offset from the beginning of the
7039 section. (Actually, \c{ELF} encodes it as the offset from the operand
7040 field of the \c{ADD} instruction, but NASM simplifies this
7041 deliberately, so you do things the same way for both \c{ELF} and
7042 \c{BSD}.) So the instruction then \e{adds} the beginning of the section,
7043 to get the real address of the GOT, and subtracts the value of
7044 \c{.get_GOT} which it knows is in \c{EBX}. Therefore, by the time
7045 that instruction has finished, \c{EBX} contains the address of the GOT.
7047 If you didn't follow that, don't worry: it's never necessary to
7048 obtain the address of the GOT by any other means, so you can put
7049 those three instructions into a macro and safely ignore them:
7056 \c add ebx,_GLOBAL_OFFSET_TABLE_+$$-%%getgot wrt ..gotpc
7060 \S{piclocal} Finding Your Local Data Items
7062 Having got the GOT, you can then use it to obtain the addresses of
7063 your data items. Most variables will reside in the sections you have
7064 declared; they can be accessed using the \I{GOTOFF
7065 relocation}\c{..gotoff} special \I\c{WRT ..gotoff}\c{WRT} type. The
7066 way this works is like this:
7068 \c lea eax,[ebx+myvar wrt ..gotoff]
7070 The expression \c{myvar wrt ..gotoff} is calculated, when the shared
7071 library is linked, to be the offset to the local variable \c{myvar}
7072 from the beginning of the GOT. Therefore, adding it to \c{EBX} as
7073 above will place the real address of \c{myvar} in \c{EAX}.
7075 If you declare variables as \c{GLOBAL} without specifying a size for
7076 them, they are shared between code modules in the library, but do
7077 not get exported from the library to the program that loaded it.
7078 They will still be in your ordinary data and BSS sections, so you
7079 can access them in the same way as local variables, using the above
7080 \c{..gotoff} mechanism.
7082 Note that due to a peculiarity of the way BSD \c{a.out} format
7083 handles this relocation type, there must be at least one non-local
7084 symbol in the same section as the address you're trying to access.
7087 \S{picextern} Finding External and Common Data Items
7089 If your library needs to get at an external variable (external to
7090 the \e{library}, not just to one of the modules within it), you must
7091 use the \I{GOT relocations}\I\c{WRT ..got}\c{..got} type to get at
7092 it. The \c{..got} type, instead of giving you the offset from the
7093 GOT base to the variable, gives you the offset from the GOT base to
7094 a GOT \e{entry} containing the address of the variable. The linker
7095 will set up this GOT entry when it builds the library, and the
7096 dynamic linker will place the correct address in it at load time. So
7097 to obtain the address of an external variable \c{extvar} in \c{EAX},
7100 \c mov eax,[ebx+extvar wrt ..got]
7102 This loads the address of \c{extvar} out of an entry in the GOT. The
7103 linker, when it builds the shared library, collects together every
7104 relocation of type \c{..got}, and builds the GOT so as to ensure it
7105 has every necessary entry present.
7107 Common variables must also be accessed in this way.
7110 \S{picglobal} Exporting Symbols to the Library User
7112 If you want to export symbols to the user of the library, you have
7113 to declare whether they are functions or data, and if they are data,
7114 you have to give the size of the data item. This is because the
7115 dynamic linker has to build \I{PLT}\i{procedure linkage table}
7116 entries for any exported functions, and also moves exported data
7117 items away from the library's data section in which they were
7120 So to export a function to users of the library, you must use
7122 \c global func:function ; declare it as a function
7128 And to export a data item such as an array, you would have to code
7130 \c global array:data array.end-array ; give the size too
7135 Be careful: If you export a variable to the library user, by
7136 declaring it as \c{GLOBAL} and supplying a size, the variable will
7137 end up living in the data section of the main program, rather than
7138 in your library's data section, where you declared it. So you will
7139 have to access your own global variable with the \c{..got} mechanism
7140 rather than \c{..gotoff}, as if it were external (which,
7141 effectively, it has become).
7143 Equally, if you need to store the address of an exported global in
7144 one of your data sections, you can't do it by means of the standard
7147 \c dataptr: dd global_data_item ; WRONG
7149 NASM will interpret this code as an ordinary relocation, in which
7150 \c{global_data_item} is merely an offset from the beginning of the
7151 \c{.data} section (or whatever); so this reference will end up
7152 pointing at your data section instead of at the exported global
7153 which resides elsewhere.
7155 Instead of the above code, then, you must write
7157 \c dataptr: dd global_data_item wrt ..sym
7159 which makes use of the special \c{WRT} type \I\c{WRT ..sym}\c{..sym}
7160 to instruct NASM to search the symbol table for a particular symbol
7161 at that address, rather than just relocating by section base.
7163 Either method will work for functions: referring to one of your
7164 functions by means of
7166 \c funcptr: dd my_function
7168 will give the user the address of the code you wrote, whereas
7170 \c funcptr: dd my_function wrt .sym
7172 will give the address of the procedure linkage table for the
7173 function, which is where the calling program will \e{believe} the
7174 function lives. Either address is a valid way to call the function.
7177 \S{picproc} Calling Procedures Outside the Library
7179 Calling procedures outside your shared library has to be done by
7180 means of a \i\e{procedure linkage table}, or \i{PLT}. The PLT is
7181 placed at a known offset from where the library is loaded, so the
7182 library code can make calls to the PLT in a position-independent
7183 way. Within the PLT there is code to jump to offsets contained in
7184 the GOT, so function calls to other shared libraries or to routines
7185 in the main program can be transparently passed off to their real
7188 To call an external routine, you must use another special PIC
7189 relocation type, \I{PLT relocations}\i\c{WRT ..plt}. This is much
7190 easier than the GOT-based ones: you simply replace calls such as
7191 \c{CALL printf} with the PLT-relative version \c{CALL printf WRT
7195 \S{link} Generating the Library File
7197 Having written some code modules and assembled them to \c{.o} files,
7198 you then generate your shared library with a command such as
7200 \c ld -shared -o library.so module1.o module2.o # for ELF
7201 \c ld -Bshareable -o library.so module1.o module2.o # for BSD
7203 For ELF, if your shared library is going to reside in system
7204 directories such as \c{/usr/lib} or \c{/lib}, it is usually worth
7205 using the \i\c{-soname} flag to the linker, to store the final
7206 library file name, with a version number, into the library:
7208 \c ld -shared -soname library.so.1 -o library.so.1.2 *.o
7210 You would then copy \c{library.so.1.2} into the library directory,
7211 and create \c{library.so.1} as a symbolic link to it.
7214 \C{mixsize} Mixing 16 and 32 Bit Code
7216 This chapter tries to cover some of the issues, largely related to
7217 unusual forms of addressing and jump instructions, encountered when
7218 writing operating system code such as protected-mode initialisation
7219 routines, which require code that operates in mixed segment sizes,
7220 such as code in a 16-bit segment trying to modify data in a 32-bit
7221 one, or jumps between different-size segments.
7224 \H{mixjump} Mixed-Size Jumps\I{jumps, mixed-size}
7226 \I{operating system, writing}\I{writing operating systems}The most
7227 common form of \i{mixed-size instruction} is the one used when
7228 writing a 32-bit OS: having done your setup in 16-bit mode, such as
7229 loading the kernel, you then have to boot it by switching into
7230 protected mode and jumping to the 32-bit kernel start address. In a
7231 fully 32-bit OS, this tends to be the \e{only} mixed-size
7232 instruction you need, since everything before it can be done in pure
7233 16-bit code, and everything after it can be pure 32-bit.
7235 This jump must specify a 48-bit far address, since the target
7236 segment is a 32-bit one. However, it must be assembled in a 16-bit
7237 segment, so just coding, for example,
7239 \c jmp 0x1234:0x56789ABC ; wrong!
7241 will not work, since the offset part of the address will be
7242 truncated to \c{0x9ABC} and the jump will be an ordinary 16-bit far
7245 The Linux kernel setup code gets round the inability of \c{as86} to
7246 generate the required instruction by coding it manually, using
7247 \c{DB} instructions. NASM can go one better than that, by actually
7248 generating the right instruction itself. Here's how to do it right:
7250 \c jmp dword 0x1234:0x56789ABC ; right
7252 \I\c{JMP DWORD}The \c{DWORD} prefix (strictly speaking, it should
7253 come \e{after} the colon, since it is declaring the \e{offset} field
7254 to be a doubleword; but NASM will accept either form, since both are
7255 unambiguous) forces the offset part to be treated as far, in the
7256 assumption that you are deliberately writing a jump from a 16-bit
7257 segment to a 32-bit one.
7259 You can do the reverse operation, jumping from a 32-bit segment to a
7260 16-bit one, by means of the \c{WORD} prefix:
7262 \c jmp word 0x8765:0x4321 ; 32 to 16 bit
7264 If the \c{WORD} prefix is specified in 16-bit mode, or the \c{DWORD}
7265 prefix in 32-bit mode, they will be ignored, since each is
7266 explicitly forcing NASM into a mode it was in anyway.
7269 \H{mixaddr} Addressing Between Different-Size Segments\I{addressing,
7270 mixed-size}\I{mixed-size addressing}
7272 If your OS is mixed 16 and 32-bit, or if you are writing a DOS
7273 extender, you are likely to have to deal with some 16-bit segments
7274 and some 32-bit ones. At some point, you will probably end up
7275 writing code in a 16-bit segment which has to access data in a
7276 32-bit segment, or vice versa.
7278 If the data you are trying to access in a 32-bit segment lies within
7279 the first 64K of the segment, you may be able to get away with using
7280 an ordinary 16-bit addressing operation for the purpose; but sooner
7281 or later, you will want to do 32-bit addressing from 16-bit mode.
7283 The easiest way to do this is to make sure you use a register for
7284 the address, since any effective address containing a 32-bit
7285 register is forced to be a 32-bit address. So you can do
7287 \c mov eax,offset_into_32_bit_segment_specified_by_fs
7288 \c mov dword [fs:eax],0x11223344
7290 This is fine, but slightly cumbersome (since it wastes an
7291 instruction and a register) if you already know the precise offset
7292 you are aiming at. The x86 architecture does allow 32-bit effective
7293 addresses to specify nothing but a 4-byte offset, so why shouldn't
7294 NASM be able to generate the best instruction for the purpose?
7296 It can. As in \k{mixjump}, you need only prefix the address with the
7297 \c{DWORD} keyword, and it will be forced to be a 32-bit address:
7299 \c mov dword [fs:dword my_offset],0x11223344
7301 Also as in \k{mixjump}, NASM is not fussy about whether the
7302 \c{DWORD} prefix comes before or after the segment override, so
7303 arguably a nicer-looking way to code the above instruction is
7305 \c mov dword [dword fs:my_offset],0x11223344
7307 Don't confuse the \c{DWORD} prefix \e{outside} the square brackets,
7308 which controls the size of the data stored at the address, with the
7309 one \c{inside} the square brackets which controls the length of the
7310 address itself. The two can quite easily be different:
7312 \c mov word [dword 0x12345678],0x9ABC
7314 This moves 16 bits of data to an address specified by a 32-bit
7317 You can also specify \c{WORD} or \c{DWORD} prefixes along with the
7318 \c{FAR} prefix to indirect far jumps or calls. For example:
7320 \c call dword far [fs:word 0x4321]
7322 This instruction contains an address specified by a 16-bit offset;
7323 it loads a 48-bit far pointer from that (16-bit segment and 32-bit
7324 offset), and calls that address.
7327 \H{mixother} Other Mixed-Size Instructions
7329 The other way you might want to access data might be using the
7330 string instructions (\c{LODSx}, \c{STOSx} and so on) or the
7331 \c{XLATB} instruction. These instructions, since they take no
7332 parameters, might seem to have no easy way to make them perform
7333 32-bit addressing when assembled in a 16-bit segment.
7335 This is the purpose of NASM's \i\c{a16}, \i\c{a32} and \i\c{a64} prefixes. If
7336 you are coding \c{LODSB} in a 16-bit segment but it is supposed to
7337 be accessing a string in a 32-bit segment, you should load the
7338 desired address into \c{ESI} and then code
7342 The prefix forces the addressing size to 32 bits, meaning that
7343 \c{LODSB} loads from \c{[DS:ESI]} instead of \c{[DS:SI]}. To access
7344 a string in a 16-bit segment when coding in a 32-bit one, the
7345 corresponding \c{a16} prefix can be used.
7347 The \c{a16}, \c{a32} and \c{a64} prefixes can be applied to any instruction
7348 in NASM's instruction table, but most of them can generate all the
7349 useful forms without them. The prefixes are necessary only for
7350 instructions with implicit addressing:
7351 \# \c{CMPSx} (\k{insCMPSB}),
7352 \# \c{SCASx} (\k{insSCASB}), \c{LODSx} (\k{insLODSB}), \c{STOSx}
7353 \# (\k{insSTOSB}), \c{MOVSx} (\k{insMOVSB}), \c{INSx} (\k{insINSB}),
7354 \# \c{OUTSx} (\k{insOUTSB}), and \c{XLATB} (\k{insXLATB}).
7355 \c{CMPSx}, \c{SCASx}, \c{LODSx}, \c{STOSx}, \c{MOVSx}, \c{INSx},
7356 \c{OUTSx}, and \c{XLATB}.
7358 various push and pop instructions (\c{PUSHA} and \c{POPF} as well as
7359 the more usual \c{PUSH} and \c{POP}) can accept \c{a16}, \c{a32} or \c{a64}
7360 prefixes to force a particular one of \c{SP}, \c{ESP} or \c{RSP} to be used
7361 as a stack pointer, in case the stack segment in use is a different
7362 size from the code segment.
7364 \c{PUSH} and \c{POP}, when applied to segment registers in 32-bit
7365 mode, also have the slightly odd behaviour that they push and pop 4
7366 bytes at a time, of which the top two are ignored and the bottom two
7367 give the value of the segment register being manipulated. To force
7368 the 16-bit behaviour of segment-register push and pop instructions,
7369 you can use the operand-size prefix \i\c{o16}:
7374 This code saves a doubleword of stack space by fitting two segment
7375 registers into the space which would normally be consumed by pushing
7378 (You can also use the \i\c{o32} prefix to force the 32-bit behaviour
7379 when in 16-bit mode, but this seems less useful.)
7382 \C{64bit} Writing 64-bit Code (Unix, Win64)
7384 This chapter attempts to cover some of the common issues involved when
7385 writing 64-bit code, to run under \i{Win64} or Unix. It covers how to
7386 write assembly code to interface with 64-bit C routines, and how to
7387 write position-independent code for shared libraries.
7389 All 64-bit code uses a flat memory model, since segmentation is not
7390 available in 64-bit mode. The one exception is the \c{FS} and \c{GS}
7391 registers, which still add their bases.
7393 Position independence in 64-bit mode is significantly simpler, since
7394 the processor supports \c{RIP}-relative addressing directly; see the
7395 \c{REL} keyword (\k{effaddr}). On most 64-bit platforms, it is
7396 probably desirable to make that the default, using the directive
7397 \c{DEFAULT REL} (\k{default}).
7399 64-bit programming is relatively similar to 32-bit programming, but
7400 of course pointers are 64 bits long; additionally, all existing
7401 platforms pass arguments in registers rather than on the stack.
7402 Furthermore, 64-bit platforms use SSE2 by default for floating point.
7403 Please see the ABI documentation for your platform.
7405 64-bit platforms differ in the sizes of the fundamental datatypes, not
7406 just from 32-bit platforms but from each other. If a specific size
7407 data type is desired, it is probably best to use the types defined in
7408 the Standard C header \c{<inttypes.h>}.
7410 In 64-bit mode, the default instruction size is still 32 bits. When
7411 loading a value into a 32-bit register (but not an 8- or 16-bit
7412 register), the upper 32 bits of the corresponding 64-bit register are
7415 \H{reg64} Register Names in 64-bit Mode
7417 NASM uses the following names for general-purpose registers in 64-bit
7418 mode, for 8-, 16-, 32- and 64-bit references, respecitively:
7420 \c AL/AH, CL/CH, DL/DH, BL/BH, SPL, BPL, SIL, DIL, R8B-R15B
7421 \c AX, CX, DX, BX, SP, BP, SI, DI, R8W-R15W
7422 \c EAX, ECX, EDX, EBX, ESP, EBP, ESI, EDI, R8D-R15D
7423 \c RAX, RCX, RDX, RBX, RSP, RBP, RSI, RDI, R8-R15
7425 This is consistent with the AMD documentation and most other
7426 assemblers. The Intel documentation, however, uses the names
7427 \c{R8L-R15L} for 8-bit references to the higher registers. It is
7428 possible to use those names by definiting them as macros; similarly,
7429 if one wants to use numeric names for the low 8 registers, define them
7430 as macros. The standard macro package \c{altreg} (see \k{pkg_altreg})
7431 can be used for this purpose.
7433 \H{id64} Immediates and Displacements in 64-bit Mode
7435 In 64-bit mode, immediates and displacements are generally only 32
7436 bits wide. NASM will therefore truncate most displacements and
7437 immediates to 32 bits.
7439 The only instruction which takes a full \i{64-bit immediate} is:
7443 NASM will produce this instruction whenever the programmer uses
7444 \c{MOV} with an immediate into a 64-bit register. If this is not
7445 desirable, simply specify the equivalent 32-bit register, which will
7446 be automatically zero-extended by the processor, or specify the
7447 immediate as \c{DWORD}:
7449 \c mov rax,foo ; 64-bit immediate
7450 \c mov rax,qword foo ; (identical)
7451 \c mov eax,foo ; 32-bit immediate, zero-extended
7452 \c mov rax,dword foo ; 32-bit immediate, sign-extended
7454 The length of these instructions are 10, 5 and 7 bytes, respectively.
7456 The only instructions which take a full \I{64-bit displacement}64-bit
7457 \e{displacement} is loading or storing, using \c{MOV}, \c{AL}, \c{AX},
7458 \c{EAX} or \c{RAX} (but no other registers) to an absolute 64-bit address.
7459 Since this is a relatively rarely used instruction (64-bit code generally uses
7460 relative addressing), the programmer has to explicitly declare the
7461 displacement size as \c{QWORD}:
7465 \c mov eax,[foo] ; 32-bit absolute disp, sign-extended
7466 \c mov eax,[a32 foo] ; 32-bit absolute disp, zero-extended
7467 \c mov eax,[qword foo] ; 64-bit absolute disp
7471 \c mov eax,[foo] ; 32-bit relative disp
7472 \c mov eax,[a32 foo] ; d:o, address truncated to 32 bits(!)
7473 \c mov eax,[qword foo] ; error
7474 \c mov eax,[abs qword foo] ; 64-bit absolute disp
7476 A sign-extended absolute displacement can access from -2 GB to +2 GB;
7477 a zero-extended absolute displacement can access from 0 to 4 GB.
7479 \H{unix64} Interfacing to 64-bit C Programs (Unix)
7481 On Unix, the 64-bit ABI is defined by the document:
7483 \W{http://www.x86-64.org/documentation/abi.pdf}\c{http://www.x86-64.org/documentation/abi.pdf}
7485 Although written for AT&T-syntax assembly, the concepts apply equally
7486 well for NASM-style assembly. What follows is a simplified summary.
7488 The first six integer arguments (from the left) are passed in \c{RDI},
7489 \c{RSI}, \c{RDX}, \c{RCX}, \c{R8}, and \c{R9}, in that order.
7490 Additional integer arguments are passed on the stack. These
7491 registers, plus \c{RAX}, \c{R10} and \c{R11} are destroyed by function
7492 calls, and thus are available for use by the function without saving.
7494 Integer return values are passed in \c{RAX} and \c{RDX}, in that order.
7496 Floating point is done using SSE registers, except for \c{long
7497 double}. Floating-point arguments are passed in \c{XMM0} to \c{XMM7};
7498 return is \c{XMM0} and \c{XMM1}. \c{long double} are passed on the
7499 stack, and returned in \c{ST0} and \c{ST1}.
7501 All SSE and x87 registers are destroyed by function calls.
7503 On 64-bit Unix, \c{long} is 64 bits.
7505 Integer and SSE register arguments are counted separately, so for the case of
7507 \c void foo(long a, double b, int c)
7509 \c{a} is passed in \c{RDI}, \c{b} in \c{XMM0}, and \c{c} in \c{ESI}.
7511 \H{win64} Interfacing to 64-bit C Programs (Win64)
7513 The Win64 ABI is described at:
7515 \W{http://msdn2.microsoft.com/en-gb/library/ms794533.aspx}\c{http://msdn2.microsoft.com/en-gb/library/ms794533.aspx}
7517 What follows is a simplified summary.
7519 The first four integer arguments are passed in \c{RCX}, \c{RDX},
7520 \c{R8} and \c{R9}, in that order. Additional integer arguments are
7521 passed on the stack. These registers, plus \c{RAX}, \c{R10} and
7522 \c{R11} are destroyed by function calls, and thus are available for
7523 use by the function without saving.
7525 Integer return values are passed in \c{RAX} only.
7527 Floating point is done using SSE registers, except for \c{long
7528 double}. Floating-point arguments are passed in \c{XMM0} to \c{XMM3};
7529 return is \c{XMM0} only.
7531 On Win64, \c{long} is 32 bits; \c{long long} or \c{_int64} is 64 bits.
7533 Integer and SSE register arguments are counted together, so for the case of
7535 \c void foo(long long a, double b, int c)
7537 \c{a} is passed in \c{RCX}, \c{b} in \c{XMM1}, and \c{c} in \c{R8D}.
7539 \C{trouble} Troubleshooting
7541 This chapter describes some of the common problems that users have
7542 been known to encounter with NASM, and answers them. It also gives
7543 instructions for reporting bugs in NASM if you find a difficulty
7544 that isn't listed here.
7547 \H{problems} Common Problems
7549 \S{inefficient} NASM Generates \i{Inefficient Code}
7551 We sometimes get `bug' reports about NASM generating inefficient, or
7552 even `wrong', code on instructions such as \c{ADD ESP,8}. This is a
7553 deliberate design feature, connected to predictability of output:
7554 NASM, on seeing \c{ADD ESP,8}, will generate the form of the
7555 instruction which leaves room for a 32-bit offset. You need to code
7556 \I\c{BYTE}\c{ADD ESP,BYTE 8} if you want the space-efficient form of
7557 the instruction. This isn't a bug, it's user error: if you prefer to
7558 have NASM produce the more efficient code automatically enable
7559 optimization with the \c{-O} option (see \k{opt-O}).
7562 \S{jmprange} My Jumps are Out of Range\I{out of range, jumps}
7564 Similarly, people complain that when they issue \i{conditional
7565 jumps} (which are \c{SHORT} by default) that try to jump too far,
7566 NASM reports `short jump out of range' instead of making the jumps
7569 This, again, is partly a predictability issue, but in fact has a
7570 more practical reason as well. NASM has no means of being told what
7571 type of processor the code it is generating will be run on; so it
7572 cannot decide for itself that it should generate \i\c{Jcc NEAR} type
7573 instructions, because it doesn't know that it's working for a 386 or
7574 above. Alternatively, it could replace the out-of-range short
7575 \c{JNE} instruction with a very short \c{JE} instruction that jumps
7576 over a \c{JMP NEAR}; this is a sensible solution for processors
7577 below a 386, but hardly efficient on processors which have good
7578 branch prediction \e{and} could have used \c{JNE NEAR} instead. So,
7579 once again, it's up to the user, not the assembler, to decide what
7580 instructions should be generated. See \k{opt-O}.
7583 \S{proborg} \i\c{ORG} Doesn't Work
7585 People writing \i{boot sector} programs in the \c{bin} format often
7586 complain that \c{ORG} doesn't work the way they'd like: in order to
7587 place the \c{0xAA55} signature word at the end of a 512-byte boot
7588 sector, people who are used to MASM tend to code
7592 \c ; some boot sector code
7597 This is not the intended use of the \c{ORG} directive in NASM, and
7598 will not work. The correct way to solve this problem in NASM is to
7599 use the \i\c{TIMES} directive, like this:
7603 \c ; some boot sector code
7605 \c TIMES 510-($-$$) DB 0
7608 The \c{TIMES} directive will insert exactly enough zero bytes into
7609 the output to move the assembly point up to 510. This method also
7610 has the advantage that if you accidentally fill your boot sector too
7611 full, NASM will catch the problem at assembly time and report it, so
7612 you won't end up with a boot sector that you have to disassemble to
7613 find out what's wrong with it.
7616 \S{probtimes} \i\c{TIMES} Doesn't Work
7618 The other common problem with the above code is people who write the
7623 by reasoning that \c{$} should be a pure number, just like 510, so
7624 the difference between them is also a pure number and can happily be
7627 NASM is a \e{modular} assembler: the various component parts are
7628 designed to be easily separable for re-use, so they don't exchange
7629 information unnecessarily. In consequence, the \c{bin} output
7630 format, even though it has been told by the \c{ORG} directive that
7631 the \c{.text} section should start at 0, does not pass that
7632 information back to the expression evaluator. So from the
7633 evaluator's point of view, \c{$} isn't a pure number: it's an offset
7634 from a section base. Therefore the difference between \c{$} and 510
7635 is also not a pure number, but involves a section base. Values
7636 involving section bases cannot be passed as arguments to \c{TIMES}.
7638 The solution, as in the previous section, is to code the \c{TIMES}
7641 \c TIMES 510-($-$$) DB 0
7643 in which \c{$} and \c{$$} are offsets from the same section base,
7644 and so their difference is a pure number. This will solve the
7645 problem and generate sensible code.
7648 \H{bugs} \i{Bugs}\I{reporting bugs}
7650 We have never yet released a version of NASM with any \e{known}
7651 bugs. That doesn't usually stop there being plenty we didn't know
7652 about, though. Any that you find should be reported firstly via the
7654 \W{https://sourceforge.net/projects/nasm/}\c{https://sourceforge.net/projects/nasm/}
7655 (click on "Bugs"), or if that fails then through one of the
7656 contacts in \k{contact}.
7658 Please read \k{qstart} first, and don't report the bug if it's
7659 listed in there as a deliberate feature. (If you think the feature
7660 is badly thought out, feel free to send us reasons why you think it
7661 should be changed, but don't just send us mail saying `This is a
7662 bug' if the documentation says we did it on purpose.) Then read
7663 \k{problems}, and don't bother reporting the bug if it's listed
7666 If you do report a bug, \e{please} give us all of the following
7669 \b What operating system you're running NASM under. DOS, Linux,
7670 NetBSD, Win16, Win32, VMS (I'd be impressed), whatever.
7672 \b If you're running NASM under DOS or Win32, tell us whether you've
7673 compiled your own executable from the DOS source archive, or whether
7674 you were using the standard distribution binaries out of the
7675 archive. If you were using a locally built executable, try to
7676 reproduce the problem using one of the standard binaries, as this
7677 will make it easier for us to reproduce your problem prior to fixing
7680 \b Which version of NASM you're using, and exactly how you invoked
7681 it. Give us the precise command line, and the contents of the
7682 \c{NASMENV} environment variable if any.
7684 \b Which versions of any supplementary programs you're using, and
7685 how you invoked them. If the problem only becomes visible at link
7686 time, tell us what linker you're using, what version of it you've
7687 got, and the exact linker command line. If the problem involves
7688 linking against object files generated by a compiler, tell us what
7689 compiler, what version, and what command line or options you used.
7690 (If you're compiling in an IDE, please try to reproduce the problem
7691 with the command-line version of the compiler.)
7693 \b If at all possible, send us a NASM source file which exhibits the
7694 problem. If this causes copyright problems (e.g. you can only
7695 reproduce the bug in restricted-distribution code) then bear in mind
7696 the following two points: firstly, we guarantee that any source code
7697 sent to us for the purposes of debugging NASM will be used \e{only}
7698 for the purposes of debugging NASM, and that we will delete all our
7699 copies of it as soon as we have found and fixed the bug or bugs in
7700 question; and secondly, we would prefer \e{not} to be mailed large
7701 chunks of code anyway. The smaller the file, the better. A
7702 three-line sample file that does nothing useful \e{except}
7703 demonstrate the problem is much easier to work with than a
7704 fully fledged ten-thousand-line program. (Of course, some errors
7705 \e{do} only crop up in large files, so this may not be possible.)
7707 \b A description of what the problem actually \e{is}. `It doesn't
7708 work' is \e{not} a helpful description! Please describe exactly what
7709 is happening that shouldn't be, or what isn't happening that should.
7710 Examples might be: `NASM generates an error message saying Line 3
7711 for an error that's actually on Line 5'; `NASM generates an error
7712 message that I believe it shouldn't be generating at all'; `NASM
7713 fails to generate an error message that I believe it \e{should} be
7714 generating'; `the object file produced from this source code crashes
7715 my linker'; `the ninth byte of the output file is 66 and I think it
7716 should be 77 instead'.
7718 \b If you believe the output file from NASM to be faulty, send it to
7719 us. That allows us to determine whether our own copy of NASM
7720 generates the same file, or whether the problem is related to
7721 portability issues between our development platforms and yours. We
7722 can handle binary files mailed to us as MIME attachments, uuencoded,
7723 and even BinHex. Alternatively, we may be able to provide an FTP
7724 site you can upload the suspect files to; but mailing them is easier
7727 \b Any other information or data files that might be helpful. If,
7728 for example, the problem involves NASM failing to generate an object
7729 file while TASM can generate an equivalent file without trouble,
7730 then send us \e{both} object files, so we can see what TASM is doing
7731 differently from us.
7734 \A{ndisasm} \i{Ndisasm}
7736 The Netwide Disassembler, NDISASM
7738 \H{ndisintro} Introduction
7741 The Netwide Disassembler is a small companion program to the Netwide
7742 Assembler, NASM. It seemed a shame to have an x86 assembler,
7743 complete with a full instruction table, and not make as much use of
7744 it as possible, so here's a disassembler which shares the
7745 instruction table (and some other bits of code) with NASM.
7747 The Netwide Disassembler does nothing except to produce
7748 disassemblies of \e{binary} source files. NDISASM does not have any
7749 understanding of object file formats, like \c{objdump}, and it will
7750 not understand \c{DOS .EXE} files like \c{debug} will. It just
7754 \H{ndisstart} Getting Started: Installation
7756 See \k{install} for installation instructions. NDISASM, like NASM,
7757 has a \c{man page} which you may want to put somewhere useful, if you
7758 are on a Unix system.
7761 \H{ndisrun} Running NDISASM
7763 To disassemble a file, you will typically use a command of the form
7765 \c ndisasm -b {16|32|64} filename
7767 NDISASM can disassemble 16-, 32- or 64-bit code equally easily,
7768 provided of course that you remember to specify which it is to work
7769 with. If no \i\c{-b} switch is present, NDISASM works in 16-bit mode
7770 by default. The \i\c{-u} switch (for USE32) also invokes 32-bit mode.
7772 Two more command line options are \i\c{-r} which reports the version
7773 number of NDISASM you are running, and \i\c{-h} which gives a short
7774 summary of command line options.
7777 \S{ndiscom} COM Files: Specifying an Origin
7779 To disassemble a \c{DOS .COM} file correctly, a disassembler must assume
7780 that the first instruction in the file is loaded at address \c{0x100},
7781 rather than at zero. NDISASM, which assumes by default that any file
7782 you give it is loaded at zero, will therefore need to be informed of
7785 The \i\c{-o} option allows you to declare a different origin for the
7786 file you are disassembling. Its argument may be expressed in any of
7787 the NASM numeric formats: decimal by default, if it begins with `\c{$}'
7788 or `\c{0x}' or ends in `\c{H}' it's \c{hex}, if it ends in `\c{Q}' it's
7789 \c{octal}, and if it ends in `\c{B}' it's \c{binary}.
7791 Hence, to disassemble a \c{.COM} file:
7793 \c ndisasm -o100h filename.com
7798 \S{ndissync} Code Following Data: Synchronisation
7800 Suppose you are disassembling a file which contains some data which
7801 isn't machine code, and \e{then} contains some machine code. NDISASM
7802 will faithfully plough through the data section, producing machine
7803 instructions wherever it can (although most of them will look
7804 bizarre, and some may have unusual prefixes, e.g. `\c{FS OR AX,0x240A}'),
7805 and generating `DB' instructions ever so often if it's totally stumped.
7806 Then it will reach the code section.
7808 Supposing NDISASM has just finished generating a strange machine
7809 instruction from part of the data section, and its file position is
7810 now one byte \e{before} the beginning of the code section. It's
7811 entirely possible that another spurious instruction will get
7812 generated, starting with the final byte of the data section, and
7813 then the correct first instruction in the code section will not be
7814 seen because the starting point skipped over it. This isn't really
7817 To avoid this, you can specify a `\i\c{synchronisation}' point, or indeed
7818 as many synchronisation points as you like (although NDISASM can
7819 only handle 2147483647 sync points internally). The definition of a sync
7820 point is this: NDISASM guarantees to hit sync points exactly during
7821 disassembly. If it is thinking about generating an instruction which
7822 would cause it to jump over a sync point, it will discard that
7823 instruction and output a `\c{db}' instead. So it \e{will} start
7824 disassembly exactly from the sync point, and so you \e{will} see all
7825 the instructions in your code section.
7827 Sync points are specified using the \i\c{-s} option: they are measured
7828 in terms of the program origin, not the file position. So if you
7829 want to synchronize after 32 bytes of a \c{.COM} file, you would have to
7832 \c ndisasm -o100h -s120h file.com
7836 \c ndisasm -o100h -s20h file.com
7838 As stated above, you can specify multiple sync markers if you need
7839 to, just by repeating the \c{-s} option.
7842 \S{ndisisync} Mixed Code and Data: Automatic (Intelligent) Synchronisation
7845 Suppose you are disassembling the boot sector of a \c{DOS} floppy (maybe
7846 it has a virus, and you need to understand the virus so that you
7847 know what kinds of damage it might have done you). Typically, this
7848 will contain a \c{JMP} instruction, then some data, then the rest of the
7849 code. So there is a very good chance of NDISASM being \e{misaligned}
7850 when the data ends and the code begins. Hence a sync point is
7853 On the other hand, why should you have to specify the sync point
7854 manually? What you'd do in order to find where the sync point would
7855 be, surely, would be to read the \c{JMP} instruction, and then to use
7856 its target address as a sync point. So can NDISASM do that for you?
7858 The answer, of course, is yes: using either of the synonymous
7859 switches \i\c{-a} (for automatic sync) or \i\c{-i} (for intelligent
7860 sync) will enable \c{auto-sync} mode. Auto-sync mode automatically
7861 generates a sync point for any forward-referring PC-relative jump or
7862 call instruction that NDISASM encounters. (Since NDISASM is one-pass,
7863 if it encounters a PC-relative jump whose target has already been
7864 processed, there isn't much it can do about it...)
7866 Only PC-relative jumps are processed, since an absolute jump is
7867 either through a register (in which case NDISASM doesn't know what
7868 the register contains) or involves a segment address (in which case
7869 the target code isn't in the same segment that NDISASM is working
7870 in, and so the sync point can't be placed anywhere useful).
7872 For some kinds of file, this mechanism will automatically put sync
7873 points in all the right places, and save you from having to place
7874 any sync points manually. However, it should be stressed that
7875 auto-sync mode is \e{not} guaranteed to catch all the sync points, and
7876 you may still have to place some manually.
7878 Auto-sync mode doesn't prevent you from declaring manual sync
7879 points: it just adds automatically generated ones to the ones you
7880 provide. It's perfectly feasible to specify \c{-i} \e{and} some \c{-s}
7883 Another caveat with auto-sync mode is that if, by some unpleasant
7884 fluke, something in your data section should disassemble to a
7885 PC-relative call or jump instruction, NDISASM may obediently place a
7886 sync point in a totally random place, for example in the middle of
7887 one of the instructions in your code section. So you may end up with
7888 a wrong disassembly even if you use auto-sync. Again, there isn't
7889 much I can do about this. If you have problems, you'll have to use
7890 manual sync points, or use the \c{-k} option (documented below) to
7891 suppress disassembly of the data area.
7894 \S{ndisother} Other Options
7896 The \i\c{-e} option skips a header on the file, by ignoring the first N
7897 bytes. This means that the header is \e{not} counted towards the
7898 disassembly offset: if you give \c{-e10 -o10}, disassembly will start
7899 at byte 10 in the file, and this will be given offset 10, not 20.
7901 The \i\c{-k} option is provided with two comma-separated numeric
7902 arguments, the first of which is an assembly offset and the second
7903 is a number of bytes to skip. This \e{will} count the skipped bytes
7904 towards the assembly offset: its use is to suppress disassembly of a
7905 data section which wouldn't contain anything you wanted to see
7909 \H{ndisbugs} Bugs and Improvements
7911 There are no known bugs. However, any you find, with patches if
7912 possible, should be sent to
7913 \W{mailto:nasm-bugs@lists.sourceforge.net}\c{nasm-bugs@lists.sourceforge.net}, or to the
7915 \W{https://sourceforge.net/projects/nasm/}\c{https://sourceforge.net/projects/nasm/}
7916 and we'll try to fix them. Feel free to send contributions and
7917 new features as well.
7919 \A{inslist} \i{Instruction List}
7921 \H{inslistintro} Introduction
7923 The following sections show the instructions which NASM currently supports. For each
7924 instruction, there is a separate entry for each supported addressing mode. The third
7925 column shows the processor type in which the instruction was introduced and,
7926 when appropriate, one or more usage flags.
7930 \A{changelog} \i{NASM Version History}