1 \# --------------------------------------------------------------------------
3 \# Copyright 1996-2012 The NASM Authors - All Rights Reserved
4 \# See the file AUTHORS included with the NASM distribution for
5 \# the specific copyright holders.
7 \# Redistribution and use in source and binary forms, with or without
8 \# modification, are permitted provided that the following
11 \# * Redistributions of source code must retain the above copyright
12 \# notice, this list of conditions and the following disclaimer.
13 \# * Redistributions in binary form must reproduce the above
14 \# copyright notice, this list of conditions and the following
15 \# disclaimer in the documentation and/or other materials provided
16 \# with the distribution.
18 \# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
19 \# CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
20 \# INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
21 \# MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
22 \# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
23 \# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 \# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
25 \# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
26 \# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 \# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
28 \# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
29 \# OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
30 \# EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
32 \# --------------------------------------------------------------------------
34 \# Source code to NASM documentation
36 \M{category}{Programming}
37 \M{title}{NASM - The Netwide Assembler}
39 \M{author}{The NASM Development Team}
40 \M{copyright_tail}{-- All Rights Reserved}
41 \M{license}{This document is redistributable under the license given in the file "LICENSE" distributed in the NASM archive.}
42 \M{summary}{This file documents NASM, the Netwide Assembler: an assembler targetting the Intel x86 series of processors, with portable source.}
45 \M{infotitle}{The Netwide Assembler for x86}
46 \M{epslogo}{nasmlogo.eps}
53 \IR{-MD} \c{-MD} option
54 \IR{-MF} \c{-MF} option
55 \IR{-MG} \c{-MG} option
56 \IR{-MP} \c{-MP} option
57 \IR{-MQ} \c{-MQ} option
58 \IR{-MT} \c{-MT} option
79 \IR{!=} \c{!=} operator
80 \IR{$, here} \c{$}, Here token
81 \IR{$, prefix} \c{$}, prefix
84 \IR{%%} \c{%%} operator
85 \IR{%+1} \c{%+1} and \c{%-1} syntax
87 \IR{%0} \c{%0} parameter count
89 \IR{&&} \c{&&} operator
91 \IR{..@} \c{..@} symbol prefix
93 \IR{//} \c{//} operator
95 \IR{<<} \c{<<} operator
96 \IR{<=} \c{<=} operator
97 \IR{<>} \c{<>} operator
99 \IR{==} \c{==} operator
100 \IR{>} \c{>} operator
101 \IR{>=} \c{>=} operator
102 \IR{>>} \c{>>} operator
103 \IR{?} \c{?} MASM syntax
104 \IR{^} \c{^} operator
105 \IR{^^} \c{^^} operator
106 \IR{|} \c{|} operator
107 \IR{||} \c{||} operator
108 \IR{~} \c{~} operator
109 \IR{%$} \c{%$} and \c{%$$} prefixes
111 \IR{+ opaddition} \c{+} operator, binary
112 \IR{+ opunary} \c{+} operator, unary
113 \IR{+ modifier} \c{+} modifier
114 \IR{- opsubtraction} \c{-} operator, binary
115 \IR{- opunary} \c{-} operator, unary
116 \IR{! opunary} \c{!} operator, unary
117 \IR{alignment, in bin sections} alignment, in \c{bin} sections
118 \IR{alignment, in elf sections} alignment, in \c{elf} sections
119 \IR{alignment, in win32 sections} alignment, in \c{win32} sections
120 \IR{alignment, of elf common variables} alignment, of \c{elf} common
122 \IR{alignment, in obj sections} alignment, in \c{obj} sections
123 \IR{a.out, bsd version} \c{a.out}, BSD version
124 \IR{a.out, linux version} \c{a.out}, Linux version
125 \IR{autoconf} Autoconf
127 \IR{bitwise and} bitwise AND
128 \IR{bitwise or} bitwise OR
129 \IR{bitwise xor} bitwise XOR
130 \IR{block ifs} block IFs
131 \IR{borland pascal} Borland, Pascal
132 \IR{borland's win32 compilers} Borland, Win32 compilers
133 \IR{braces, after % sign} braces, after \c{%} sign
135 \IR{c calling convention} C calling convention
136 \IR{c symbol names} C symbol names
137 \IA{critical expressions}{critical expression}
138 \IA{command line}{command-line}
139 \IA{case sensitivity}{case sensitive}
140 \IA{case-sensitive}{case sensitive}
141 \IA{case-insensitive}{case sensitive}
142 \IA{character constants}{character constant}
143 \IR{common object file format} Common Object File Format
144 \IR{common variables, alignment in elf} common variables, alignment
146 \IR{common, elf extensions to} \c{COMMON}, \c{elf} extensions to
147 \IR{common, obj extensions to} \c{COMMON}, \c{obj} extensions to
148 \IR{declaring structure} declaring structures
149 \IR{default-wrt mechanism} default-\c{WRT} mechanism
152 \IR{dll symbols, exporting} DLL symbols, exporting
153 \IR{dll symbols, importing} DLL symbols, importing
155 \IR{dos archive} DOS archive
156 \IR{dos source archive} DOS source archive
157 \IA{effective address}{effective addresses}
158 \IA{effective-address}{effective addresses}
160 \IR{elf, 16-bit code and} ELF, 16-bit code and
161 \IR{elf shared libraries} ELF, shared libraries
164 \IR{elfx32} \c{elfx32}
165 \IR{executable and linkable format} Executable and Linkable Format
166 \IR{extern, obj extensions to} \c{EXTERN}, \c{obj} extensions to
167 \IR{extern, rdf extensions to} \c{EXTERN}, \c{rdf} extensions to
168 \IR{floating-point, constants} floating-point, constants
169 \IR{floating-point, packed bcd constants} floating-point, packed BCD constants
171 \IR{freelink} FreeLink
172 \IR{functions, c calling convention} functions, C calling convention
173 \IR{functions, pascal calling convention} functions, Pascal calling
175 \IR{global, aoutb extensions to} \c{GLOBAL}, \c{aoutb} extensions to
176 \IR{global, elf extensions to} \c{GLOBAL}, \c{elf} extensions to
177 \IR{global, rdf extensions to} \c{GLOBAL}, \c{rdf} extensions to
179 \IR{got relocations} \c{GOT} relocations
180 \IR{gotoff relocation} \c{GOTOFF} relocations
181 \IR{gotpc relocation} \c{GOTPC} relocations
182 \IR{intel number formats} Intel number formats
183 \IR{linux, elf} Linux, ELF
184 \IR{linux, a.out} Linux, \c{a.out}
185 \IR{linux, as86} Linux, \c{as86}
186 \IR{logical and} logical AND
187 \IR{logical or} logical OR
188 \IR{logical xor} logical XOR
189 \IR{mach object file format} Mach, object file format
191 \IR{macho32} \c{macho32}
192 \IR{macho64} \c{macho64}
195 \IA{memory reference}{memory references}
197 \IA{misc directory}{misc subdirectory}
198 \IR{misc subdirectory} \c{misc} subdirectory
199 \IR{microsoft omf} Microsoft OMF
200 \IR{mmx registers} MMX registers
201 \IA{modr/m}{modr/m byte}
202 \IR{modr/m byte} ModR/M byte
204 \IR{ms-dos device drivers} MS-DOS device drivers
205 \IR{multipush} \c{multipush} macro
207 \IR{nasm version} NASM version
211 \IR{operating system} operating system
213 \IR{pascal calling convention}Pascal calling convention
214 \IR{passes} passes, assembly
219 \IR{plt} \c{PLT} relocations
220 \IA{pre-defining macros}{pre-define}
221 \IA{preprocessor expressions}{preprocessor, expressions}
222 \IA{preprocessor loops}{preprocessor, loops}
223 \IA{preprocessor variables}{preprocessor, variables}
224 \IA{rdoff subdirectory}{rdoff}
225 \IR{rdoff} \c{rdoff} subdirectory
226 \IR{relocatable dynamic object file format} Relocatable Dynamic
228 \IR{relocations, pic-specific} relocations, PIC-specific
229 \IA{repeating}{repeating code}
230 \IR{section alignment, in elf} section alignment, in \c{elf}
231 \IR{section alignment, in bin} section alignment, in \c{bin}
232 \IR{section alignment, in obj} section alignment, in \c{obj}
233 \IR{section alignment, in win32} section alignment, in \c{win32}
234 \IR{section, elf extensions to} \c{SECTION}, \c{elf} extensions to
235 \IR{section, win32 extensions to} \c{SECTION}, \c{win32} extensions to
236 \IR{segment alignment, in bin} segment alignment, in \c{bin}
237 \IR{segment alignment, in obj} segment alignment, in \c{obj}
238 \IR{segment, obj extensions to} \c{SEGMENT}, \c{elf} extensions to
239 \IR{segment names, borland pascal} segment names, Borland Pascal
240 \IR{shift command} \c{shift} command
242 \IR{sib byte} SIB byte
243 \IR{align, smart} \c{ALIGN}, smart
244 \IA{sectalign}{sectalign}
245 \IR{solaris x86} Solaris x86
246 \IA{standard section names}{standardized section names}
247 \IR{symbols, exporting from dlls} symbols, exporting from DLLs
248 \IR{symbols, importing from dlls} symbols, importing from DLLs
249 \IR{test subdirectory} \c{test} subdirectory
251 \IR{underscore, in c symbols} underscore, in C symbols
257 \IA{sco unix}{unix, sco}
258 \IR{unix, sco} Unix, SCO
259 \IA{unix source archive}{unix, source archive}
260 \IR{unix, source archive} Unix, source archive
261 \IA{unix system v}{unix, system v}
262 \IR{unix, system v} Unix, System V
263 \IR{unixware} UnixWare
265 \IR{version number of nasm} version number of NASM
266 \IR{visual c++} Visual C++
267 \IR{www page} WWW page
271 \IR{windows 95} Windows 95
272 \IR{windows nt} Windows NT
273 \# \IC{program entry point}{entry point, program}
274 \# \IC{program entry point}{start point, program}
275 \# \IC{MS-DOS device drivers}{device drivers, MS-DOS}
276 \# \IC{16-bit mode, versus 32-bit mode}{32-bit mode, versus 16-bit mode}
277 \# \IC{c symbol names}{symbol names, in C}
280 \C{intro} Introduction
282 \H{whatsnasm} What Is NASM?
284 The Netwide Assembler, NASM, is an 80x86 and x86-64 assembler designed
285 for portability and modularity. It supports a range of object file
286 formats, including Linux and \c{*BSD} \c{a.out}, \c{ELF}, \c{COFF},
287 \c{Mach-O}, Microsoft 16-bit \c{OBJ}, \c{Win32} and \c{Win64}. It will
288 also output plain binary files. Its syntax is designed to be simple
289 and easy to understand, similar to Intel's but less complex. It
290 supports all currently known x86 architectural extensions, and has
291 strong support for macros.
294 \S{yaasm} Why Yet Another Assembler?
296 The Netwide Assembler grew out of an idea on \i\c{comp.lang.asm.x86}
297 (or possibly \i\c{alt.lang.asm} - I forget which), which was
298 essentially that there didn't seem to be a good \e{free} x86-series
299 assembler around, and that maybe someone ought to write one.
301 \b \i\c{a86} is good, but not free, and in particular you don't get any
302 32-bit capability until you pay. It's DOS only, too.
304 \b \i\c{gas} is free, and ports over to DOS and Unix, but it's not
305 very good, since it's designed to be a back end to \i\c{gcc}, which
306 always feeds it correct code. So its error checking is minimal. Also,
307 its syntax is horrible, from the point of view of anyone trying to
308 actually \e{write} anything in it. Plus you can't write 16-bit code in
311 \b \i\c{as86} is specific to Minix and Linux, and (my version at least)
312 doesn't seem to have much (or any) documentation.
314 \b \i\c{MASM} isn't very good, and it's (was) expensive, and it runs only under
317 \b \i\c{TASM} is better, but still strives for MASM compatibility,
318 which means millions of directives and tons of red tape. And its syntax
319 is essentially MASM's, with the contradictions and quirks that
320 entails (although it sorts out some of those by means of Ideal mode.)
321 It's expensive too. And it's DOS-only.
323 So here, for your coding pleasure, is NASM. At present it's
324 still in prototype stage - we don't promise that it can outperform
325 any of these assemblers. But please, \e{please} send us bug reports,
326 fixes, helpful information, and anything else you can get your hands
327 on (and thanks to the many people who've done this already! You all
328 know who you are), and we'll improve it out of all recognition.
332 \S{legal} \i{License} Conditions
334 Please see the file \c{LICENSE}, supplied as part of any NASM
335 distribution archive, for the license conditions under which you may
336 use NASM. NASM is now under the so-called 2-clause BSD license, also
337 known as the simplified BSD license.
339 Copyright 1996-2011 the NASM Authors - All rights reserved.
341 Redistribution and use in source and binary forms, with or without
342 modification, are permitted provided that the following conditions are
345 \b Redistributions of source code must retain the above copyright
346 notice, this list of conditions and the following disclaimer.
348 \b Redistributions in binary form must reproduce the above copyright
349 notice, this list of conditions and the following disclaimer in the
350 documentation and/or other materials provided with the distribution.
352 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
353 CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
354 INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
355 MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
356 DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
357 CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
358 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
359 NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
360 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
361 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
362 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
363 OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
364 EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
367 \H{contact} Contact Information
369 The current version of NASM (since about 0.98.08) is maintained by a
370 team of developers, accessible through the \c{nasm-devel} mailing list
371 (see below for the link).
372 If you want to report a bug, please read \k{bugs} first.
374 NASM has a \i{website} at
375 \W{http://www.nasm.us/}\c{http://www.nasm.us/}. If it's not there,
378 \i{New releases}, \i{release candidates}, and \I{snapshots, daily
379 development}\i{daily development snapshots} of NASM are available from
380 the official web site.
382 Announcements are posted to
383 \W{news:comp.lang.asm.x86}\i\c{comp.lang.asm.x86},
385 \W{http://www.freshmeat.net/}\c{http://www.freshmeat.net/}.
387 If you want information about the current development status, please
388 subscribe to the \i\c{nasm-devel} email list; see link from the
392 \H{install} Installation
394 \S{instdos} \i{Installing} NASM under MS-\i{DOS} or Windows
396 Once you've obtained the appropriate archive for NASM,
397 \i\c{nasm-XXX-dos.zip} or \i\c{nasm-XXX-win32.zip} (where \c{XXX}
398 denotes the version number of NASM contained in the archive), unpack
399 it into its own directory (for example \c{c:\\nasm}).
401 The archive will contain a set of executable files: the NASM
402 executable file \i\c{nasm.exe}, the NDISASM executable file
403 \i\c{ndisasm.exe}, and possibly additional utilities to handle the
406 The only file NASM needs to run is its own executable, so copy
407 \c{nasm.exe} to a directory on your PATH, or alternatively edit
408 \i\c{autoexec.bat} to add the \c{nasm} directory to your
409 \i\c{PATH} (to do that under Windows XP, go to Start > Control Panel >
410 System > Advanced > Environment Variables; these instructions may work
411 under other versions of Windows as well.)
413 That's it - NASM is installed. You don't need the nasm directory
414 to be present to run NASM (unless you've added it to your \c{PATH}),
415 so you can delete it if you need to save space; however, you may
416 want to keep the documentation or test programs.
418 If you've downloaded the \i{DOS source archive}, \i\c{nasm-XXX.zip},
419 the \c{nasm} directory will also contain the full NASM \i{source
420 code}, and a selection of \i{Makefiles} you can (hopefully) use to
421 rebuild your copy of NASM from scratch. See the file \c{INSTALL} in
424 Note that a number of files are generated from other files by Perl
425 scripts. Although the NASM source distribution includes these
426 generated files, you will need to rebuild them (and hence, will need a
427 Perl interpreter) if you change insns.dat, standard.mac or the
428 documentation. It is possible future source distributions may not
429 include these files at all. Ports of \i{Perl} for a variety of
430 platforms, including DOS and Windows, are available from
431 \W{http://www.cpan.org/ports/}\i{www.cpan.org}.
434 \S{instdos} Installing NASM under \i{Unix}
436 Once you've obtained the \i{Unix source archive} for NASM,
437 \i\c{nasm-XXX.tar.gz} (where \c{XXX} denotes the version number of
438 NASM contained in the archive), unpack it into a directory such
439 as \c{/usr/local/src}. The archive, when unpacked, will create its
440 own subdirectory \c{nasm-XXX}.
442 NASM is an \I{Autoconf}\I\c{configure}auto-configuring package: once
443 you've unpacked it, \c{cd} to the directory it's been unpacked into
444 and type \c{./configure}. This shell script will find the best C
445 compiler to use for building NASM and set up \i{Makefiles}
448 Once NASM has auto-configured, you can type \i\c{make} to build the
449 \c{nasm} and \c{ndisasm} binaries, and then \c{make install} to
450 install them in \c{/usr/local/bin} and install the \i{man pages}
451 \i\c{nasm.1} and \i\c{ndisasm.1} in \c{/usr/local/man/man1}.
452 Alternatively, you can give options such as \c{--prefix} to the
453 configure script (see the file \i\c{INSTALL} for more details), or
454 install the programs yourself.
456 NASM also comes with a set of utilities for handling the \c{RDOFF}
457 custom object-file format, which are in the \i\c{rdoff} subdirectory
458 of the NASM archive. You can build these with \c{make rdf} and
459 install them with \c{make rdf_install}, if you want them.
462 \C{running} Running NASM
464 \H{syntax} NASM \i{Command-Line} Syntax
466 To assemble a file, you issue a command of the form
468 \c nasm -f <format> <filename> [-o <output>]
472 \c nasm -f elf myfile.asm
474 will assemble \c{myfile.asm} into an \c{ELF} object file \c{myfile.o}. And
476 \c nasm -f bin myfile.asm -o myfile.com
478 will assemble \c{myfile.asm} into a raw binary file \c{myfile.com}.
480 To produce a listing file, with the hex codes output from NASM
481 displayed on the left of the original sources, use the \c{-l} option
482 to give a listing file name, for example:
484 \c nasm -f coff myfile.asm -l myfile.lst
486 To get further usage instructions from NASM, try typing
490 As \c{-hf}, this will also list the available output file formats, and what they
493 If you use Linux but aren't sure whether your system is \c{a.out}
498 (in the directory in which you put the NASM binary when you
499 installed it). If it says something like
501 \c nasm: ELF 32-bit LSB executable i386 (386 and up) Version 1
503 then your system is \c{ELF}, and you should use the option \c{-f elf}
504 when you want NASM to produce Linux object files. If it says
506 \c nasm: Linux/i386 demand-paged executable (QMAGIC)
508 or something similar, your system is \c{a.out}, and you should use
509 \c{-f aout} instead (Linux \c{a.out} systems have long been obsolete,
510 and are rare these days.)
512 Like Unix compilers and assemblers, NASM is silent unless it
513 goes wrong: you won't see any output at all, unless it gives error
517 \S{opt-o} The \i\c{-o} Option: Specifying the Output File Name
519 NASM will normally choose the name of your output file for you;
520 precisely how it does this is dependent on the object file format.
521 For Microsoft object file formats (\c{obj}, \c{win32} and \c{win64}),
522 it will remove the \c{.asm} \i{extension} (or whatever extension you
523 like to use - NASM doesn't care) from your source file name and
524 substitute \c{.obj}. For Unix object file formats (\c{aout}, \c{as86},
525 \c{coff}, \c{elf32}, \c{elf64}, \c{elfx32}, \c{ieee}, \c{macho32} and
526 \c{macho64}) it will substitute \c{.o}. For \c{dbg}, \c{rdf}, \c{ith}
527 and \c{srec}, it will use \c{.dbg}, \c{.rdf}, \c{.ith} and \c{.srec},
528 respectively, and for the \c{bin} format it will simply remove the
529 extension, so that \c{myfile.asm} produces the output file \c{myfile}.
531 If the output file already exists, NASM will overwrite it, unless it
532 has the same name as the input file, in which case it will give a
533 warning and use \i\c{nasm.out} as the output file name instead.
535 For situations in which this behaviour is unacceptable, NASM
536 provides the \c{-o} command-line option, which allows you to specify
537 your desired output file name. You invoke \c{-o} by following it
538 with the name you wish for the output file, either with or without
539 an intervening space. For example:
541 \c nasm -f bin program.asm -o program.com
542 \c nasm -f bin driver.asm -odriver.sys
544 Note that this is a small o, and is different from a capital O , which
545 is used to specify the number of optimisation passes required. See \k{opt-O}.
548 \S{opt-f} The \i\c{-f} Option: Specifying the \i{Output File Format}
550 If you do not supply the \c{-f} option to NASM, it will choose an
551 output file format for you itself. In the distribution versions of
552 NASM, the default is always \i\c{bin}; if you've compiled your own
553 copy of NASM, you can redefine \i\c{OF_DEFAULT} at compile time and
554 choose what you want the default to be.
556 Like \c{-o}, the intervening space between \c{-f} and the output
557 file format is optional; so \c{-f elf} and \c{-felf} are both valid.
559 A complete list of the available output file formats can be given by
560 issuing the command \i\c{nasm -hf}.
563 \S{opt-l} The \i\c{-l} Option: Generating a \i{Listing File}
565 If you supply the \c{-l} option to NASM, followed (with the usual
566 optional space) by a file name, NASM will generate a
567 \i{source-listing file} for you, in which addresses and generated
568 code are listed on the left, and the actual source code, with
569 expansions of multi-line macros (except those which specifically
570 request no expansion in source listings: see \k{nolist}) on the
573 \c nasm -f elf myfile.asm -l myfile.lst
575 If a list file is selected, you may turn off listing for a
576 section of your source with \c{[list -]}, and turn it back on
577 with \c{[list +]}, (the default, obviously). There is no "user
578 form" (without the brackets). This can be used to list only
579 sections of interest, avoiding excessively long listings.
582 \S{opt-M} The \i\c{-M} Option: Generate \i{Makefile Dependencies}
584 This option can be used to generate makefile dependencies on stdout.
585 This can be redirected to a file for further processing. For example:
587 \c nasm -M myfile.asm > myfile.dep
590 \S{opt-MG} The \i\c{-MG} Option: Generate \i{Makefile Dependencies}
592 This option can be used to generate makefile dependencies on stdout.
593 This differs from the \c{-M} option in that if a nonexisting file is
594 encountered, it is assumed to be a generated file and is added to the
595 dependency list without a prefix.
598 \S{opt-MF} The \i\c\{-MF} Option: Set Makefile Dependency File
600 This option can be used with the \c{-M} or \c{-MG} options to send the
601 output to a file, rather than to stdout. For example:
603 \c nasm -M -MF myfile.dep myfile.asm
606 \S{opt-MD} The \i\c{-MD} Option: Assemble and Generate Dependencies
608 The \c{-MD} option acts as the combination of the \c{-M} and \c{-MF}
609 options (i.e. a filename has to be specified.) However, unlike the
610 \c{-M} or \c{-MG} options, \c{-MD} does \e{not} inhibit the normal
611 operation of the assembler. Use this to automatically generate
612 updated dependencies with every assembly session. For example:
614 \c nasm -f elf -o myfile.o -MD myfile.dep myfile.asm
617 \S{opt-MT} The \i\c{-MT} Option: Dependency Target Name
619 The \c{-MT} option can be used to override the default name of the
620 dependency target. This is normally the same as the output filename,
621 specified by the \c{-o} option.
624 \S{opt-MQ} The \i\c{-MQ} Option: Dependency Target Name (Quoted)
626 The \c{-MQ} option acts as the \c{-MT} option, except it tries to
627 quote characters that have special meaning in Makefile syntax. This
628 is not foolproof, as not all characters with special meaning are
632 \S{opt-MP} The \i\c{-MP} Option: Emit phony targets
634 When used with any of the dependency generation options, the \c{-MP}
635 option causes NASM to emit a phony target without dependencies for
636 each header file. This prevents Make from complaining if a header
637 file has been removed.
640 \S{opt-F} The \i\c{-F} Option: Selecting a \i{Debug Information Format}
642 This option is used to select the format of the debug information
643 emitted into the output file, to be used by a debugger (or \e{will}
644 be). Prior to version 2.03.01, the use of this switch did \e{not} enable
645 output of the selected debug info format. Use \c{-g}, see \k{opt-g},
646 to enable output. Versions 2.03.01 and later automatically enable \c{-g}
647 if \c{-F} is specified.
649 A complete list of the available debug file formats for an output
650 format can be seen by issuing the command \c{nasm -f <format> -y}. Not
651 all output formats currently support debugging output. See \k{opt-y}.
653 This should not be confused with the \c{-f dbg} output format option which
654 is not built into NASM by default. For information on how
655 to enable it when building from the sources, see \k{dbgfmt}.
658 \S{opt-g} The \i\c{-g} Option: Enabling \i{Debug Information}.
660 This option can be used to generate debugging information in the specified
661 format. See \k{opt-F}. Using \c{-g} without \c{-F} results in emitting
662 debug info in the default format, if any, for the selected output format.
663 If no debug information is currently implemented in the selected output
664 format, \c{-g} is \e{silently ignored}.
667 \S{opt-X} The \i\c{-X} Option: Selecting an \i{Error Reporting Format}
669 This option can be used to select an error reporting format for any
670 error messages that might be produced by NASM.
672 Currently, two error reporting formats may be selected. They are
673 the \c{-Xvc} option and the \c{-Xgnu} option. The GNU format is
674 the default and looks like this:
676 \c filename.asm:65: error: specific error message
678 where \c{filename.asm} is the name of the source file in which the
679 error was detected, \c{65} is the source file line number on which
680 the error was detected, \c{error} is the severity of the error (this
681 could be \c{warning}), and \c{specific error message} is a more
682 detailed text message which should help pinpoint the exact problem.
684 The other format, specified by \c{-Xvc} is the style used by Microsoft
685 Visual C++ and some other programs. It looks like this:
687 \c filename.asm(65) : error: specific error message
689 where the only difference is that the line number is in parentheses
690 instead of being delimited by colons.
692 See also the \c{Visual C++} output format, \k{win32fmt}.
694 \S{opt-Z} The \i\c{-Z} Option: Send Errors to a File
696 Under \I{DOS}\c{MS-DOS} it can be difficult (though there are ways) to
697 redirect the standard-error output of a program to a file. Since
698 NASM usually produces its warning and \i{error messages} on
699 \i\c{stderr}, this can make it hard to capture the errors if (for
700 example) you want to load them into an editor.
702 NASM therefore provides the \c{-Z} option, taking a filename argument
703 which causes errors to be sent to the specified files rather than
704 standard error. Therefore you can \I{redirecting errors}redirect
705 the errors into a file by typing
707 \c nasm -Z myfile.err -f obj myfile.asm
709 In earlier versions of NASM, this option was called \c{-E}, but it was
710 changed since \c{-E} is an option conventionally used for
711 preprocessing only, with disastrous results. See \k{opt-E}.
713 \S{opt-s} The \i\c{-s} Option: Send Errors to \i\c{stdout}
715 The \c{-s} option redirects \i{error messages} to \c{stdout} rather
716 than \c{stderr}, so it can be redirected under \I{DOS}\c{MS-DOS}. To
717 assemble the file \c{myfile.asm} and pipe its output to the \c{more}
718 program, you can type:
720 \c nasm -s -f obj myfile.asm | more
722 See also the \c{-Z} option, \k{opt-Z}.
725 \S{opt-i} The \i\c{-i}\I\c{-I} Option: Include File Search Directories
727 When NASM sees the \i\c{%include} or \i\c{%pathsearch} directive in a
728 source file (see \k{include}, \k{pathsearch} or \k{incbin}), it will
729 search for the given file not only in the current directory, but also
730 in any directories specified on the command line by the use of the
731 \c{-i} option. Therefore you can include files from a \i{macro
732 library}, for example, by typing
734 \c nasm -ic:\macrolib\ -f obj myfile.asm
736 (As usual, a space between \c{-i} and the path name is allowed, and
739 NASM, in the interests of complete source-code portability, does not
740 understand the file naming conventions of the OS it is running on;
741 the string you provide as an argument to the \c{-i} option will be
742 prepended exactly as written to the name of the include file.
743 Therefore the trailing backslash in the above example is necessary.
744 Under Unix, a trailing forward slash is similarly necessary.
746 (You can use this to your advantage, if you're really \i{perverse},
747 by noting that the option \c{-ifoo} will cause \c{%include "bar.i"}
748 to search for the file \c{foobar.i}...)
750 If you want to define a \e{standard} \i{include search path},
751 similar to \c{/usr/include} on Unix systems, you should place one or
752 more \c{-i} directives in the \c{NASMENV} environment variable (see
755 For Makefile compatibility with many C compilers, this option can also
756 be specified as \c{-I}.
759 \S{opt-p} The \i\c{-p}\I\c{-P} Option: \I{pre-including files}Pre-Include a File
761 \I\c{%include}NASM allows you to specify files to be
762 \e{pre-included} into your source file, by the use of the \c{-p}
765 \c nasm myfile.asm -p myinc.inc
767 is equivalent to running \c{nasm myfile.asm} and placing the
768 directive \c{%include "myinc.inc"} at the start of the file.
770 For consistency with the \c{-I}, \c{-D} and \c{-U} options, this
771 option can also be specified as \c{-P}.
774 \S{opt-d} The \i\c{-d}\I\c{-D} Option: \I{pre-defining macros}Pre-Define a Macro
776 \I\c{%define}Just as the \c{-p} option gives an alternative to placing
777 \c{%include} directives at the start of a source file, the \c{-d}
778 option gives an alternative to placing a \c{%define} directive. You
781 \c nasm myfile.asm -dFOO=100
783 as an alternative to placing the directive
787 at the start of the file. You can miss off the macro value, as well:
788 the option \c{-dFOO} is equivalent to coding \c{%define FOO}. This
789 form of the directive may be useful for selecting \i{assembly-time
790 options} which are then tested using \c{%ifdef}, for example
793 For Makefile compatibility with many C compilers, this option can also
794 be specified as \c{-D}.
797 \S{opt-u} The \i\c{-u}\I\c{-U} Option: \I{Undefining macros}Undefine a Macro
799 \I\c{%undef}The \c{-u} option undefines a macro that would otherwise
800 have been pre-defined, either automatically or by a \c{-p} or \c{-d}
801 option specified earlier on the command lines.
803 For example, the following command line:
805 \c nasm myfile.asm -dFOO=100 -uFOO
807 would result in \c{FOO} \e{not} being a predefined macro in the
808 program. This is useful to override options specified at a different
811 For Makefile compatibility with many C compilers, this option can also
812 be specified as \c{-U}.
815 \S{opt-E} The \i\c{-E}\I{-e} Option: Preprocess Only
817 NASM allows the \i{preprocessor} to be run on its own, up to a
818 point. Using the \c{-E} option (which requires no arguments) will
819 cause NASM to preprocess its input file, expand all the macro
820 references, remove all the comments and preprocessor directives, and
821 print the resulting file on standard output (or save it to a file,
822 if the \c{-o} option is also used).
824 This option cannot be applied to programs which require the
825 preprocessor to evaluate \I{preprocessor expressions}\i{expressions}
826 which depend on the values of symbols: so code such as
828 \c %assign tablesize ($-tablestart)
830 will cause an error in \i{preprocess-only mode}.
832 For compatiblity with older version of NASM, this option can also be
833 written \c{-e}. \c{-E} in older versions of NASM was the equivalent
834 of the current \c{-Z} option, \k{opt-Z}.
836 \S{opt-a} The \i\c{-a} Option: Don't Preprocess At All
838 If NASM is being used as the back end to a compiler, it might be
839 desirable to \I{suppressing preprocessing}suppress preprocessing
840 completely and assume the compiler has already done it, to save time
841 and increase compilation speeds. The \c{-a} option, requiring no
842 argument, instructs NASM to replace its powerful \i{preprocessor}
843 with a \i{stub preprocessor} which does nothing.
846 \S{opt-O} The \i\c{-O} Option: Specifying \i{Multipass Optimization}
848 Using the \c{-O} option, you can tell NASM to carry out different
849 levels of optimization. The syntax is:
851 \b \c{-O0}: No optimization. All operands take their long forms,
852 if a short form is not specified, except conditional jumps.
853 This is intended to match NASM 0.98 behavior.
855 \b \c{-O1}: Minimal optimization. As above, but immediate operands
856 which will fit in a signed byte are optimized,
857 unless the long form is specified. Conditional jumps default
858 to the long form unless otherwise specified.
860 \b \c{-Ox} (where \c{x} is the actual letter \c{x}): Multipass optimization.
861 Minimize branch offsets and signed immediate bytes,
862 overriding size specification unless the \c{strict} keyword
863 has been used (see \k{strict}). For compatibility with earlier
864 releases, the letter \c{x} may also be any number greater than
865 one. This number has no effect on the actual number of passes.
867 The \c{-Ox} mode is recommended for most uses, and is the default
870 Note that this is a capital \c{O}, and is different from a small \c{o}, which
871 is used to specify the output file name. See \k{opt-o}.
874 \S{opt-t} The \i\c{-t} Option: Enable TASM Compatibility Mode
876 NASM includes a limited form of compatibility with Borland's \i\c{TASM}.
877 When NASM's \c{-t} option is used, the following changes are made:
879 \b local labels may be prefixed with \c{@@} instead of \c{.}
881 \b size override is supported within brackets. In TASM compatible mode,
882 a size override inside square brackets changes the size of the operand,
883 and not the address type of the operand as it does in NASM syntax. E.g.
884 \c{mov eax,[DWORD val]} is valid syntax in TASM compatibility mode.
885 Note that you lose the ability to override the default address type for
888 \b unprefixed forms of some directives supported (\c{arg}, \c{elif},
889 \c{else}, \c{endif}, \c{if}, \c{ifdef}, \c{ifdifi}, \c{ifndef},
890 \c{include}, \c{local})
892 \S{opt-w} The \i\c{-w} and \i\c{-W} Options: Enable or Disable Assembly \i{Warnings}
894 NASM can observe many conditions during the course of assembly which
895 are worth mentioning to the user, but not a sufficiently severe
896 error to justify NASM refusing to generate an output file. These
897 conditions are reported like errors, but come up with the word
898 `warning' before the message. Warnings do not prevent NASM from
899 generating an output file and returning a success status to the
902 Some conditions are even less severe than that: they are only
903 sometimes worth mentioning to the user. Therefore NASM supports the
904 \c{-w} command-line option, which enables or disables certain
905 classes of assembly warning. Such warning classes are described by a
906 name, for example \c{orphan-labels}; you can enable warnings of
907 this class by the command-line option \c{-w+orphan-labels} and
908 disable it by \c{-w-orphan-labels}.
910 The \i{suppressible warning} classes are:
912 \b \i\c{macro-params} covers warnings about \i{multi-line macros}
913 being invoked with the wrong number of parameters. This warning
914 class is enabled by default; see \k{mlmacover} for an example of why
915 you might want to disable it.
917 \b \i\c{macro-selfref} warns if a macro references itself. This
918 warning class is disabled by default.
920 \b\i\c{macro-defaults} warns when a macro has more default
921 parameters than optional parameters. This warning class
922 is enabled by default; see \k{mlmacdef} for why you might want to disable it.
924 \b \i\c{orphan-labels} covers warnings about source lines which
925 contain no instruction but define a label without a trailing colon.
926 NASM warns about this somewhat obscure condition by default;
927 see \k{syntax} for more information.
929 \b \i\c{number-overflow} covers warnings about numeric constants which
930 don't fit in 64 bits. This warning class is enabled by default.
932 \b \i\c{gnu-elf-extensions} warns if 8-bit or 16-bit relocations
933 are used in \c{-f elf} format. The GNU extensions allow this.
934 This warning class is disabled by default.
936 \b \i\c{float-overflow} warns about floating point overflow.
939 \b \i\c{float-denorm} warns about floating point denormals.
942 \b \i\c{float-underflow} warns about floating point underflow.
945 \b \i\c{float-toolong} warns about too many digits in floating-point numbers.
948 \b \i\c{user} controls \c{%warning} directives (see \k{pperror}).
951 \b \i\c{error} causes warnings to be treated as errors. Disabled by
954 \b \i\c{all} is an alias for \e{all} suppressible warning classes (not
955 including \c{error}). Thus, \c{-w+all} enables all available warnings.
957 In addition, you can set warning classes across sections.
958 Warning classes may be enabled with \i\c{[warning +warning-name]},
959 disabled with \i\c{[warning -warning-name]} or reset to their
960 original value with \i\c{[warning *warning-name]}. No "user form"
961 (without the brackets) exists.
963 Since version 2.00, NASM has also supported the gcc-like syntax
964 \c{-Wwarning} and \c{-Wno-warning} instead of \c{-w+warning} and
965 \c{-w-warning}, respectively.
968 \S{opt-v} The \i\c{-v} Option: Display \i{Version} Info
970 Typing \c{NASM -v} will display the version of NASM which you are using,
971 and the date on which it was compiled.
973 You will need the version number if you report a bug.
975 \S{opt-y} The \i\c{-y} Option: Display Available Debug Info Formats
977 Typing \c{nasm -f <option> -y} will display a list of the available
978 debug info formats for the given output format. The default format
979 is indicated by an asterisk. For example:
983 \c valid debug formats for 'elf32' output format are
984 \c ('*' denotes default):
985 \c * stabs ELF32 (i386) stabs debug format for Linux
986 \c dwarf elf32 (i386) dwarf debug format for Linux
989 \S{opt-pfix} The \i\c{--prefix} and \i\c{--postfix} Options.
991 The \c{--prefix} and \c{--postfix} options prepend or append
992 (respectively) the given argument to all \c{global} or
993 \c{extern} variables. E.g. \c{--prefix _} will prepend the
994 underscore to all global and external variables, as C sometimes
995 (but not always) likes it.
998 \S{nasmenv} The \i\c{NASMENV} \i{Environment} Variable
1000 If you define an environment variable called \c{NASMENV}, the program
1001 will interpret it as a list of extra command-line options, which are
1002 processed before the real command line. You can use this to define
1003 standard search directories for include files, by putting \c{-i}
1004 options in the \c{NASMENV} variable.
1006 The value of the variable is split up at white space, so that the
1007 value \c{-s -ic:\\nasmlib\\} will be treated as two separate options.
1008 However, that means that the value \c{-dNAME="my name"} won't do
1009 what you might want, because it will be split at the space and the
1010 NASM command-line processing will get confused by the two
1011 nonsensical words \c{-dNAME="my} and \c{name"}.
1013 To get round this, NASM provides a feature whereby, if you begin the
1014 \c{NASMENV} environment variable with some character that isn't a minus
1015 sign, then NASM will treat this character as the \i{separator
1016 character} for options. So setting the \c{NASMENV} variable to the
1017 value \c{!-s!-ic:\\nasmlib\\} is equivalent to setting it to \c{-s
1018 -ic:\\nasmlib\\}, but \c{!-dNAME="my name"} will work.
1020 This environment variable was previously called \c{NASM}. This was
1021 changed with version 0.98.31.
1024 \H{qstart} \i{Quick Start} for \i{MASM} Users
1026 If you're used to writing programs with MASM, or with \i{TASM} in
1027 MASM-compatible (non-Ideal) mode, or with \i\c{a86}, this section
1028 attempts to outline the major differences between MASM's syntax and
1029 NASM's. If you're not already used to MASM, it's probably worth
1030 skipping this section.
1033 \S{qscs} NASM Is \I{case sensitivity}Case-Sensitive
1035 One simple difference is that NASM is case-sensitive. It makes a
1036 difference whether you call your label \c{foo}, \c{Foo} or \c{FOO}.
1037 If you're assembling to \c{DOS} or \c{OS/2} \c{.OBJ} files, you can
1038 invoke the \i\c{UPPERCASE} directive (documented in \k{objfmt}) to
1039 ensure that all symbols exported to other code modules are forced
1040 to be upper case; but even then, \e{within} a single module, NASM
1041 will distinguish between labels differing only in case.
1044 \S{qsbrackets} NASM Requires \i{Square Brackets} For \i{Memory References}
1046 NASM was designed with simplicity of syntax in mind. One of the
1047 \i{design goals} of NASM is that it should be possible, as far as is
1048 practical, for the user to look at a single line of NASM code
1049 and tell what opcode is generated by it. You can't do this in MASM:
1050 if you declare, for example,
1055 then the two lines of code
1060 generate completely different opcodes, despite having
1061 identical-looking syntaxes.
1063 NASM avoids this undesirable situation by having a much simpler
1064 syntax for memory references. The rule is simply that any access to
1065 the \e{contents} of a memory location requires square brackets
1066 around the address, and any access to the \e{address} of a variable
1067 doesn't. So an instruction of the form \c{mov ax,foo} will
1068 \e{always} refer to a compile-time constant, whether it's an \c{EQU}
1069 or the address of a variable; and to access the \e{contents} of the
1070 variable \c{bar}, you must code \c{mov ax,[bar]}.
1072 This also means that NASM has no need for MASM's \i\c{OFFSET}
1073 keyword, since the MASM code \c{mov ax,offset bar} means exactly the
1074 same thing as NASM's \c{mov ax,bar}. If you're trying to get
1075 large amounts of MASM code to assemble sensibly under NASM, you
1076 can always code \c{%idefine offset} to make the preprocessor treat
1077 the \c{OFFSET} keyword as a no-op.
1079 This issue is even more confusing in \i\c{a86}, where declaring a
1080 label with a trailing colon defines it to be a `label' as opposed to
1081 a `variable' and causes \c{a86} to adopt NASM-style semantics; so in
1082 \c{a86}, \c{mov ax,var} has different behaviour depending on whether
1083 \c{var} was declared as \c{var: dw 0} (a label) or \c{var dw 0} (a
1084 word-size variable). NASM is very simple by comparison:
1085 \e{everything} is a label.
1087 NASM, in the interests of simplicity, also does not support the
1088 \i{hybrid syntaxes} supported by MASM and its clones, such as
1089 \c{mov ax,table[bx]}, where a memory reference is denoted by one
1090 portion outside square brackets and another portion inside. The
1091 correct syntax for the above is \c{mov ax,[table+bx]}. Likewise,
1092 \c{mov ax,es:[di]} is wrong and \c{mov ax,[es:di]} is right.
1095 \S{qstypes} NASM Doesn't Store \i{Variable Types}
1097 NASM, by design, chooses not to remember the types of variables you
1098 declare. Whereas MASM will remember, on seeing \c{var dw 0}, that
1099 you declared \c{var} as a word-size variable, and will then be able
1100 to fill in the \i{ambiguity} in the size of the instruction \c{mov
1101 var,2}, NASM will deliberately remember nothing about the symbol
1102 \c{var} except where it begins, and so you must explicitly code
1103 \c{mov word [var],2}.
1105 For this reason, NASM doesn't support the \c{LODS}, \c{MOVS},
1106 \c{STOS}, \c{SCAS}, \c{CMPS}, \c{INS}, or \c{OUTS} instructions,
1107 but only supports the forms such as \c{LODSB}, \c{MOVSW}, and
1108 \c{SCASD}, which explicitly specify the size of the components of
1109 the strings being manipulated.
1112 \S{qsassume} NASM Doesn't \i\c{ASSUME}
1114 As part of NASM's drive for simplicity, it also does not support the
1115 \c{ASSUME} directive. NASM will not keep track of what values you
1116 choose to put in your segment registers, and will never
1117 \e{automatically} generate a \i{segment override} prefix.
1120 \S{qsmodel} NASM Doesn't Support \i{Memory Models}
1122 NASM also does not have any directives to support different 16-bit
1123 memory models. The programmer has to keep track of which functions
1124 are supposed to be called with a \i{far call} and which with a
1125 \i{near call}, and is responsible for putting the correct form of
1126 \c{RET} instruction (\c{RETN} or \c{RETF}; NASM accepts \c{RET}
1127 itself as an alternate form for \c{RETN}); in addition, the
1128 programmer is responsible for coding CALL FAR instructions where
1129 necessary when calling \e{external} functions, and must also keep
1130 track of which external variable definitions are far and which are
1134 \S{qsfpu} \i{Floating-Point} Differences
1136 NASM uses different names to refer to floating-point registers from
1137 MASM: where MASM would call them \c{ST(0)}, \c{ST(1)} and so on, and
1138 \i\c{a86} would call them simply \c{0}, \c{1} and so on, NASM
1139 chooses to call them \c{st0}, \c{st1} etc.
1141 As of version 0.96, NASM now treats the instructions with
1142 \i{`nowait'} forms in the same way as MASM-compatible assemblers.
1143 The idiosyncratic treatment employed by 0.95 and earlier was based
1144 on a misunderstanding by the authors.
1147 \S{qsother} Other Differences
1149 For historical reasons, NASM uses the keyword \i\c{TWORD} where MASM
1150 and compatible assemblers use \i\c{TBYTE}.
1152 NASM does not declare \i{uninitialized storage} in the same way as
1153 MASM: where a MASM programmer might use \c{stack db 64 dup (?)},
1154 NASM requires \c{stack resb 64}, intended to be read as `reserve 64
1155 bytes'. For a limited amount of compatibility, since NASM treats
1156 \c{?} as a valid character in symbol names, you can code \c{? equ 0}
1157 and then writing \c{dw ?} will at least do something vaguely useful.
1158 \I\c{RESB}\i\c{DUP} is still not a supported syntax, however.
1160 In addition to all of this, macros and directives work completely
1161 differently to MASM. See \k{preproc} and \k{directive} for further
1165 \C{lang} The NASM Language
1167 \H{syntax} Layout of a NASM Source Line
1169 Like most assemblers, each NASM source line contains (unless it
1170 is a macro, a preprocessor directive or an assembler directive: see
1171 \k{preproc} and \k{directive}) some combination of the four fields
1173 \c label: instruction operands ; comment
1175 As usual, most of these fields are optional; the presence or absence
1176 of any combination of a label, an instruction and a comment is allowed.
1177 Of course, the operand field is either required or forbidden by the
1178 presence and nature of the instruction field.
1180 NASM uses backslash (\\) as the line continuation character; if a line
1181 ends with backslash, the next line is considered to be a part of the
1182 backslash-ended line.
1184 NASM places no restrictions on white space within a line: labels may
1185 have white space before them, or instructions may have no space
1186 before them, or anything. The \i{colon} after a label is also
1187 optional. (Note that this means that if you intend to code \c{lodsb}
1188 alone on a line, and type \c{lodab} by accident, then that's still a
1189 valid source line which does nothing but define a label. Running
1190 NASM with the command-line option
1191 \I{orphan-labels}\c{-w+orphan-labels} will cause it to warn you if
1192 you define a label alone on a line without a \i{trailing colon}.)
1194 \i{Valid characters} in labels are letters, numbers, \c{_}, \c{$},
1195 \c{#}, \c{@}, \c{~}, \c{.}, and \c{?}. The only characters which may
1196 be used as the \e{first} character of an identifier are letters,
1197 \c{.} (with special meaning: see \k{locallab}), \c{_} and \c{?}.
1198 An identifier may also be prefixed with a \I{$, prefix}\c{$} to
1199 indicate that it is intended to be read as an identifier and not a
1200 reserved word; thus, if some other module you are linking with
1201 defines a symbol called \c{eax}, you can refer to \c{$eax} in NASM
1202 code to distinguish the symbol from the register. Maximum length of
1203 an identifier is 4095 characters.
1205 The instruction field may contain any machine instruction: Pentium
1206 and P6 instructions, FPU instructions, MMX instructions and even
1207 undocumented instructions are all supported. The instruction may be
1208 prefixed by \c{LOCK}, \c{REP}, \c{REPE}/\c{REPZ} or
1209 \c{REPNE}/\c{REPNZ}, in the usual way. Explicit \I{address-size
1210 prefixes}address-size and \i{operand-size prefixes} \i\c{A16},
1211 \i\c{A32}, \i\c{A64}, \i\c{O16} and \i\c{O32}, \i\c{O64} are provided - one example of their use
1212 is given in \k{mixsize}. You can also use the name of a \I{segment
1213 override}segment register as an instruction prefix: coding
1214 \c{es mov [bx],ax} is equivalent to coding \c{mov [es:bx],ax}. We
1215 recommend the latter syntax, since it is consistent with other
1216 syntactic features of the language, but for instructions such as
1217 \c{LODSB}, which has no operands and yet can require a segment
1218 override, there is no clean syntactic way to proceed apart from
1221 An instruction is not required to use a prefix: prefixes such as
1222 \c{CS}, \c{A32}, \c{LOCK} or \c{REPE} can appear on a line by
1223 themselves, and NASM will just generate the prefix bytes.
1225 In addition to actual machine instructions, NASM also supports a
1226 number of pseudo-instructions, described in \k{pseudop}.
1228 Instruction \i{operands} may take a number of forms: they can be
1229 registers, described simply by the register name (e.g. \c{ax},
1230 \c{bp}, \c{ebx}, \c{cr0}: NASM does not use the \c{gas}-style
1231 syntax in which register names must be prefixed by a \c{%} sign), or
1232 they can be \i{effective addresses} (see \k{effaddr}), constants
1233 (\k{const}) or expressions (\k{expr}).
1235 For x87 \i{floating-point} instructions, NASM accepts a wide range of
1236 syntaxes: you can use two-operand forms like MASM supports, or you
1237 can use NASM's native single-operand forms in most cases.
1239 \# all forms of each supported instruction are given in
1241 For example, you can code:
1243 \c fadd st1 ; this sets st0 := st0 + st1
1244 \c fadd st0,st1 ; so does this
1246 \c fadd st1,st0 ; this sets st1 := st1 + st0
1247 \c fadd to st1 ; so does this
1249 Almost any x87 floating-point instruction that references memory must
1250 use one of the prefixes \i\c{DWORD}, \i\c{QWORD} or \i\c{TWORD} to
1251 indicate what size of \i{memory operand} it refers to.
1254 \H{pseudop} \i{Pseudo-Instructions}
1256 Pseudo-instructions are things which, though not real x86 machine
1257 instructions, are used in the instruction field anyway because that's
1258 the most convenient place to put them. The current pseudo-instructions
1259 are \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, \i\c{DO} and
1260 \i\c{DY}; their \i{uninitialized} counterparts \i\c{RESB}, \i\c{RESW},
1261 \i\c{RESD}, \i\c{RESQ}, \i\c{REST}, \i\c{RESO} and \i\c{RESY}; the
1262 \i\c{INCBIN} command, the \i\c{EQU} command, and the \i\c{TIMES}
1266 \S{db} \c{DB} and Friends: Declaring Initialized Data
1268 \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, \i\c{DO} and
1269 \i\c{DY} are used, much as in MASM, to declare initialized data in the
1270 output file. They can be invoked in a wide range of ways:
1271 \I{floating-point}\I{character constant}\I{string constant}
1273 \c db 0x55 ; just the byte 0x55
1274 \c db 0x55,0x56,0x57 ; three bytes in succession
1275 \c db 'a',0x55 ; character constants are OK
1276 \c db 'hello',13,10,'$' ; so are string constants
1277 \c dw 0x1234 ; 0x34 0x12
1278 \c dw 'a' ; 0x61 0x00 (it's just a number)
1279 \c dw 'ab' ; 0x61 0x62 (character constant)
1280 \c dw 'abc' ; 0x61 0x62 0x63 0x00 (string)
1281 \c dd 0x12345678 ; 0x78 0x56 0x34 0x12
1282 \c dd 1.234567e20 ; floating-point constant
1283 \c dq 0x123456789abcdef0 ; eight byte constant
1284 \c dq 1.234567e20 ; double-precision float
1285 \c dt 1.234567e20 ; extended-precision float
1287 \c{DT}, \c{DO} and \c{DY} do not accept \i{numeric constants} as operands.
1290 \S{resb} \c{RESB} and Friends: Declaring \i{Uninitialized} Data
1292 \i\c{RESB}, \i\c{RESW}, \i\c{RESD}, \i\c{RESQ}, \i\c{REST}, \i\c{RESO}
1293 and \i\c{RESY} are designed to be used in the BSS section of a module:
1294 they declare \e{uninitialized} storage space. Each takes a single
1295 operand, which is the number of bytes, words, doublewords or whatever
1296 to reserve. As stated in \k{qsother}, NASM does not support the
1297 MASM/TASM syntax of reserving uninitialized space by writing
1298 \I\c{?}\c{DW ?} or similar things: this is what it does instead. The
1299 operand to a \c{RESB}-type pseudo-instruction is a \i\e{critical
1300 expression}: see \k{crit}.
1304 \c buffer: resb 64 ; reserve 64 bytes
1305 \c wordvar: resw 1 ; reserve a word
1306 \c realarray resq 10 ; array of ten reals
1307 \c ymmval: resy 1 ; one YMM register
1309 \S{incbin} \i\c{INCBIN}: Including External \i{Binary Files}
1311 \c{INCBIN} is borrowed from the old Amiga assembler \i{DevPac}: it
1312 includes a binary file verbatim into the output file. This can be
1313 handy for (for example) including \i{graphics} and \i{sound} data
1314 directly into a game executable file. It can be called in one of
1317 \c incbin "file.dat" ; include the whole file
1318 \c incbin "file.dat",1024 ; skip the first 1024 bytes
1319 \c incbin "file.dat",1024,512 ; skip the first 1024, and
1320 \c ; actually include at most 512
1322 \c{INCBIN} is both a directive and a standard macro; the standard
1323 macro version searches for the file in the include file search path
1324 and adds the file to the dependency lists. This macro can be
1325 overridden if desired.
1328 \S{equ} \i\c{EQU}: Defining Constants
1330 \c{EQU} defines a symbol to a given constant value: when \c{EQU} is
1331 used, the source line must contain a label. The action of \c{EQU} is
1332 to define the given label name to the value of its (only) operand.
1333 This definition is absolute, and cannot change later. So, for
1336 \c message db 'hello, world'
1337 \c msglen equ $-message
1339 defines \c{msglen} to be the constant 12. \c{msglen} may not then be
1340 redefined later. This is not a \i{preprocessor} definition either:
1341 the value of \c{msglen} is evaluated \e{once}, using the value of
1342 \c{$} (see \k{expr} for an explanation of \c{$}) at the point of
1343 definition, rather than being evaluated wherever it is referenced
1344 and using the value of \c{$} at the point of reference.
1347 \S{times} \i\c{TIMES}: \i{Repeating} Instructions or Data
1349 The \c{TIMES} prefix causes the instruction to be assembled multiple
1350 times. This is partly present as NASM's equivalent of the \i\c{DUP}
1351 syntax supported by \i{MASM}-compatible assemblers, in that you can
1354 \c zerobuf: times 64 db 0
1356 or similar things; but \c{TIMES} is more versatile than that. The
1357 argument to \c{TIMES} is not just a numeric constant, but a numeric
1358 \e{expression}, so you can do things like
1360 \c buffer: db 'hello, world'
1361 \c times 64-$+buffer db ' '
1363 which will store exactly enough spaces to make the total length of
1364 \c{buffer} up to 64. Finally, \c{TIMES} can be applied to ordinary
1365 instructions, so you can code trivial \i{unrolled loops} in it:
1369 Note that there is no effective difference between \c{times 100 resb
1370 1} and \c{resb 100}, except that the latter will be assembled about
1371 100 times faster due to the internal structure of the assembler.
1373 The operand to \c{TIMES} is a critical expression (\k{crit}).
1375 Note also that \c{TIMES} can't be applied to \i{macros}: the reason
1376 for this is that \c{TIMES} is processed after the macro phase, which
1377 allows the argument to \c{TIMES} to contain expressions such as
1378 \c{64-$+buffer} as above. To repeat more than one line of code, or a
1379 complex macro, use the preprocessor \i\c{%rep} directive.
1382 \H{effaddr} Effective Addresses
1384 An \i{effective address} is any operand to an instruction which
1385 \I{memory reference}references memory. Effective addresses, in NASM,
1386 have a very simple syntax: they consist of an expression evaluating
1387 to the desired address, enclosed in \i{square brackets}. For
1392 \c mov ax,[wordvar+1]
1393 \c mov ax,[es:wordvar+bx]
1395 Anything not conforming to this simple system is not a valid memory
1396 reference in NASM, for example \c{es:wordvar[bx]}.
1398 More complicated effective addresses, such as those involving more
1399 than one register, work in exactly the same way:
1401 \c mov eax,[ebx*2+ecx+offset]
1404 NASM is capable of doing \i{algebra} on these effective addresses,
1405 so that things which don't necessarily \e{look} legal are perfectly
1408 \c mov eax,[ebx*5] ; assembles as [ebx*4+ebx]
1409 \c mov eax,[label1*2-label2] ; ie [label1+(label1-label2)]
1411 Some forms of effective address have more than one assembled form;
1412 in most such cases NASM will generate the smallest form it can. For
1413 example, there are distinct assembled forms for the 32-bit effective
1414 addresses \c{[eax*2+0]} and \c{[eax+eax]}, and NASM will generally
1415 generate the latter on the grounds that the former requires four
1416 bytes to store a zero offset.
1418 NASM has a hinting mechanism which will cause \c{[eax+ebx]} and
1419 \c{[ebx+eax]} to generate different opcodes; this is occasionally
1420 useful because \c{[esi+ebp]} and \c{[ebp+esi]} have different
1421 default segment registers.
1423 However, you can force NASM to generate an effective address in a
1424 particular form by the use of the keywords \c{BYTE}, \c{WORD},
1425 \c{DWORD} and \c{NOSPLIT}. If you need \c{[eax+3]} to be assembled
1426 using a double-word offset field instead of the one byte NASM will
1427 normally generate, you can code \c{[dword eax+3]}. Similarly, you
1428 can force NASM to use a byte offset for a small value which it
1429 hasn't seen on the first pass (see \k{crit} for an example of such a
1430 code fragment) by using \c{[byte eax+offset]}. As special cases,
1431 \c{[byte eax]} will code \c{[eax+0]} with a byte offset of zero, and
1432 \c{[dword eax]} will code it with a double-word offset of zero. The
1433 normal form, \c{[eax]}, will be coded with no offset field.
1435 The form described in the previous paragraph is also useful if you
1436 are trying to access data in a 32-bit segment from within 16 bit code.
1437 For more information on this see the section on mixed-size addressing
1438 (\k{mixaddr}). In particular, if you need to access data with a known
1439 offset that is larger than will fit in a 16-bit value, if you don't
1440 specify that it is a dword offset, nasm will cause the high word of
1441 the offset to be lost.
1443 Similarly, NASM will split \c{[eax*2]} into \c{[eax+eax]} because
1444 that allows the offset field to be absent and space to be saved; in
1445 fact, it will also split \c{[eax*2+offset]} into
1446 \c{[eax+eax+offset]}. You can combat this behaviour by the use of
1447 the \c{NOSPLIT} keyword: \c{[nosplit eax*2]} will force
1448 \c{[eax*2+0]} to be generated literally.
1450 In 64-bit mode, NASM will by default generate absolute addresses. The
1451 \i\c{REL} keyword makes it produce \c{RIP}-relative addresses. Since
1452 this is frequently the normally desired behaviour, see the \c{DEFAULT}
1453 directive (\k{default}). The keyword \i\c{ABS} overrides \i\c{REL}.
1456 \H{const} \i{Constants}
1458 NASM understands four different types of constant: numeric,
1459 character, string and floating-point.
1462 \S{numconst} \i{Numeric Constants}
1464 A numeric constant is simply a number. NASM allows you to specify
1465 numbers in a variety of number bases, in a variety of ways: you can
1466 suffix \c{H} or \c{X}, \c{D} or \c{T}, \c{Q} or \c{O}, and \c{B} or
1467 \c{Y} for \i{hexadecimal}, \i{decimal}, \i{octal} and \i{binary}
1468 respectively, or you can prefix \c{0x}, for hexadecimal in the style
1469 of C, or you can prefix \c{$} for hexadecimal in the style of Borland
1470 Pascal or Motorola Assemblers. Note, though, that the \I{$,
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{0d} or
1475 \c{0t} for decimal, \c{0o} or \c{0q} for octal, and \c{0b} or \c{0y}
1476 for binary. Please note that unlike C, a \c{0} prefix by itself does
1477 \e{not} imply an octal constant!
1479 Numeric constants can have underscores (\c{_}) interspersed to break
1482 Some examples (all producing exactly the same code):
1484 \c mov ax,200 ; decimal
1485 \c mov ax,0200 ; still decimal
1486 \c mov ax,0200d ; explicitly decimal
1487 \c mov ax,0d200 ; also decimal
1488 \c mov ax,0c8h ; hex
1489 \c mov ax,$0c8 ; hex again: the 0 is required
1490 \c mov ax,0xc8 ; hex yet again
1491 \c mov ax,0hc8 ; still hex
1492 \c mov ax,310q ; octal
1493 \c mov ax,310o ; octal again
1494 \c mov ax,0o310 ; octal yet again
1495 \c mov ax,0q310 ; octal yet again
1496 \c mov ax,11001000b ; binary
1497 \c mov ax,1100_1000b ; same binary constant
1498 \c mov ax,1100_1000y ; same binary constant once more
1499 \c mov ax,0b1100_1000 ; same binary constant yet again
1500 \c mov ax,0y1100_1000 ; same binary constant yet again
1502 \S{strings} \I{Strings}\i{Character Strings}
1504 A character string consists of up to eight characters enclosed in
1505 either single quotes (\c{'...'}), double quotes (\c{"..."}) or
1506 backquotes (\c{`...`}). Single or double quotes are equivalent to
1507 NASM (except of course that surrounding the constant with single
1508 quotes allows double quotes to appear within it and vice versa); the
1509 contents of those are represented verbatim. Strings enclosed in
1510 backquotes support C-style \c{\\}-escapes for special characters.
1513 The following \i{escape sequences} are recognized by backquoted strings:
1515 \c \' single quote (')
1516 \c \" double quote (")
1518 \c \\\ backslash (\)
1519 \c \? question mark (?)
1527 \c \e ESC (ASCII 27)
1528 \c \377 Up to 3 octal digits - literal byte
1529 \c \xFF Up to 2 hexadecimal digits - literal byte
1530 \c \u1234 4 hexadecimal digits - Unicode character
1531 \c \U12345678 8 hexadecimal digits - Unicode character
1533 All other escape sequences are reserved. Note that \c{\\0}, meaning a
1534 \c{NUL} character (ASCII 0), is a special case of the octal escape
1537 \i{Unicode} characters specified with \c{\\u} or \c{\\U} are converted to
1538 \i{UTF-8}. For example, the following lines are all equivalent:
1540 \c db `\u263a` ; UTF-8 smiley face
1541 \c db `\xe2\x98\xba` ; UTF-8 smiley face
1542 \c db 0E2h, 098h, 0BAh ; UTF-8 smiley face
1545 \S{chrconst} \i{Character Constants}
1547 A character constant consists of a string up to eight bytes long, used
1548 in an expression context. It is treated as if it was an integer.
1550 A character constant with more than one byte will be arranged
1551 with \i{little-endian} order in mind: if you code
1555 then the constant generated is not \c{0x61626364}, but
1556 \c{0x64636261}, so that if you were then to store the value into
1557 memory, it would read \c{abcd} rather than \c{dcba}. This is also
1558 the sense of character constants understood by the Pentium's
1559 \i\c{CPUID} instruction.
1562 \S{strconst} \i{String Constants}
1564 String constants are character strings used in the context of some
1565 pseudo-instructions, namely the
1566 \I\c{DW}\I\c{DD}\I\c{DQ}\I\c{DT}\I\c{DO}\I\c{DY}\i\c{DB} family and
1567 \i\c{INCBIN} (where it represents a filename.) They are also used in
1568 certain preprocessor directives.
1570 A string constant looks like a character constant, only longer. It
1571 is treated as a concatenation of maximum-size character constants
1572 for the conditions. So the following are equivalent:
1574 \c db 'hello' ; string constant
1575 \c db 'h','e','l','l','o' ; equivalent character constants
1577 And the following are also equivalent:
1579 \c dd 'ninechars' ; doubleword string constant
1580 \c dd 'nine','char','s' ; becomes three doublewords
1581 \c db 'ninechars',0,0,0 ; and really looks like this
1583 Note that when used in a string-supporting context, quoted strings are
1584 treated as a string constants even if they are short enough to be a
1585 character constant, because otherwise \c{db 'ab'} would have the same
1586 effect as \c{db 'a'}, which would be silly. Similarly, three-character
1587 or four-character constants are treated as strings when they are
1588 operands to \c{DW}, and so forth.
1590 \S{unicode} \I{UTF-16}\I{UTF-32}\i{Unicode} Strings
1592 The special operators \i\c{__utf16__} and \i\c{__utf32__} allows
1593 definition of Unicode strings. They take a string in UTF-8 format and
1594 converts it to (littleendian) UTF-16 or UTF-32, respectively.
1598 \c %define u(x) __utf16__(x)
1599 \c %define w(x) __utf32__(x)
1601 \c dw u('C:\WINDOWS'), 0 ; Pathname in UTF-16
1602 \c dd w(`A + B = \u206a`), 0 ; String in UTF-32
1604 \c{__utf16__} and \c{__utf32__} can be applied either to strings
1605 passed to the \c{DB} family instructions, or to character constants in
1606 an expression context.
1608 \S{fltconst} \I{floating-point, constants}Floating-Point Constants
1610 \i{Floating-point} constants are acceptable only as arguments to
1611 \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, and \i\c{DO}, or as
1612 arguments to the special operators \i\c{__float8__},
1613 \i\c{__float16__}, \i\c{__float32__}, \i\c{__float64__},
1614 \i\c{__float80m__}, \i\c{__float80e__}, \i\c{__float128l__}, and
1615 \i\c{__float128h__}.
1617 Floating-point constants are expressed in the traditional form:
1618 digits, then a period, then optionally more digits, then optionally an
1619 \c{E} followed by an exponent. The period is mandatory, so that NASM
1620 can distinguish between \c{dd 1}, which declares an integer constant,
1621 and \c{dd 1.0} which declares a floating-point constant.
1623 NASM also support C99-style hexadecimal floating-point: \c{0x},
1624 hexadecimal digits, period, optionally more hexadeximal digits, then
1625 optionally a \c{P} followed by a \e{binary} (not hexadecimal) exponent
1626 in decimal notation. As an extension, NASM additionally supports the
1627 \c{0h} and \c{$} prefixes for hexadecimal, as well binary and octal
1628 floating-point, using the \c{0b} or \c{0y} and \c{0o} or \c{0q}
1629 prefixes, respectively.
1631 Underscores to break up groups of digits are permitted in
1632 floating-point constants as well.
1636 \c db -0.2 ; "Quarter precision"
1637 \c dw -0.5 ; IEEE 754r/SSE5 half precision
1638 \c dd 1.2 ; an easy one
1639 \c dd 1.222_222_222 ; underscores are permitted
1640 \c dd 0x1p+2 ; 1.0x2^2 = 4.0
1641 \c dq 0x1p+32 ; 1.0x2^32 = 4 294 967 296.0
1642 \c dq 1.e10 ; 10 000 000 000.0
1643 \c dq 1.e+10 ; synonymous with 1.e10
1644 \c dq 1.e-10 ; 0.000 000 000 1
1645 \c dt 3.141592653589793238462 ; pi
1646 \c do 1.e+4000 ; IEEE 754r quad precision
1648 The 8-bit "quarter-precision" floating-point format is
1649 sign:exponent:mantissa = 1:4:3 with an exponent bias of 7. This
1650 appears to be the most frequently used 8-bit floating-point format,
1651 although it is not covered by any formal standard. This is sometimes
1652 called a "\i{minifloat}."
1654 The special operators are used to produce floating-point numbers in
1655 other contexts. They produce the binary representation of a specific
1656 floating-point number as an integer, and can use anywhere integer
1657 constants are used in an expression. \c{__float80m__} and
1658 \c{__float80e__} produce the 64-bit mantissa and 16-bit exponent of an
1659 80-bit floating-point number, and \c{__float128l__} and
1660 \c{__float128h__} produce the lower and upper 64-bit halves of a 128-bit
1661 floating-point number, respectively.
1665 \c mov rax,__float64__(3.141592653589793238462)
1667 ... would assign the binary representation of pi as a 64-bit floating
1668 point number into \c{RAX}. This is exactly equivalent to:
1670 \c mov rax,0x400921fb54442d18
1672 NASM cannot do compile-time arithmetic on floating-point constants.
1673 This is because NASM is designed to be portable - although it always
1674 generates code to run on x86 processors, the assembler itself can
1675 run on any system with an ANSI C compiler. Therefore, the assembler
1676 cannot guarantee the presence of a floating-point unit capable of
1677 handling the \i{Intel number formats}, and so for NASM to be able to
1678 do floating arithmetic it would have to include its own complete set
1679 of floating-point routines, which would significantly increase the
1680 size of the assembler for very little benefit.
1682 The special tokens \i\c{__Infinity__}, \i\c{__QNaN__} (or
1683 \i\c{__NaN__}) and \i\c{__SNaN__} can be used to generate
1684 \I{infinity}infinities, quiet \i{NaN}s, and signalling NaNs,
1685 respectively. These are normally used as macros:
1687 \c %define Inf __Infinity__
1688 \c %define NaN __QNaN__
1690 \c dq +1.5, -Inf, NaN ; Double-precision constants
1692 The \c{%use fp} standard macro package contains a set of convenience
1693 macros. See \k{pkg_fp}.
1695 \S{bcdconst} \I{floating-point, packed BCD constants}Packed BCD Constants
1697 x87-style packed BCD constants can be used in the same contexts as
1698 80-bit floating-point numbers. They are suffixed with \c{p} or
1699 prefixed with \c{0p}, and can include up to 18 decimal digits.
1701 As with other numeric constants, underscores can be used to separate
1706 \c dt 12_345_678_901_245_678p
1707 \c dt -12_345_678_901_245_678p
1712 \H{expr} \i{Expressions}
1714 Expressions in NASM are similar in syntax to those in C. Expressions
1715 are evaluated as 64-bit integers which are then adjusted to the
1718 NASM supports two special tokens in expressions, allowing
1719 calculations to involve the current assembly position: the
1720 \I{$, here}\c{$} and \i\c{$$} tokens. \c{$} evaluates to the assembly
1721 position at the beginning of the line containing the expression; so
1722 you can code an \i{infinite loop} using \c{JMP $}. \c{$$} evaluates
1723 to the beginning of the current section; so you can tell how far
1724 into the section you are by using \c{($-$$)}.
1726 The arithmetic \i{operators} provided by NASM are listed here, in
1727 increasing order of \i{precedence}.
1730 \S{expor} \i\c{|}: \i{Bitwise OR} Operator
1732 The \c{|} operator gives a bitwise OR, exactly as performed by the
1733 \c{OR} machine instruction. Bitwise OR is the lowest-priority
1734 arithmetic operator supported by NASM.
1737 \S{expxor} \i\c{^}: \i{Bitwise XOR} Operator
1739 \c{^} provides the bitwise XOR operation.
1742 \S{expand} \i\c{&}: \i{Bitwise AND} Operator
1744 \c{&} provides the bitwise AND operation.
1747 \S{expshift} \i\c{<<} and \i\c{>>}: \i{Bit Shift} Operators
1749 \c{<<} gives a bit-shift to the left, just as it does in C. So \c{5<<3}
1750 evaluates to 5 times 8, or 40. \c{>>} gives a bit-shift to the
1751 right; in NASM, such a shift is \e{always} unsigned, so that
1752 the bits shifted in from the left-hand end are filled with zero
1753 rather than a sign-extension of the previous highest bit.
1756 \S{expplmi} \I{+ opaddition}\c{+} and \I{- opsubtraction}\c{-}:
1757 \i{Addition} and \i{Subtraction} Operators
1759 The \c{+} and \c{-} operators do perfectly ordinary addition and
1763 \S{expmul} \i\c{*}, \i\c{/}, \i\c{//}, \i\c{%} and \i\c{%%}:
1764 \i{Multiplication} and \i{Division}
1766 \c{*} is the multiplication operator. \c{/} and \c{//} are both
1767 division operators: \c{/} is \i{unsigned division} and \c{//} is
1768 \i{signed division}. Similarly, \c{%} and \c{%%} provide \I{unsigned
1769 modulo}\I{modulo operators}unsigned and
1770 \i{signed modulo} operators respectively.
1772 NASM, like ANSI C, provides no guarantees about the sensible
1773 operation of the signed modulo operator.
1775 Since the \c{%} character is used extensively by the macro
1776 \i{preprocessor}, you should ensure that both the signed and unsigned
1777 modulo operators are followed by white space wherever they appear.
1780 \S{expmul} \i{Unary Operators}: \I{+ opunary}\c{+}, \I{- opunary}\c{-},
1781 \i\c{~}, \I{! opunary}\c{!} and \i\c{SEG}
1783 The highest-priority operators in NASM's expression grammar are
1784 those which only apply to one argument. \c{-} negates its operand,
1785 \c{+} does nothing (it's provided for symmetry with \c{-}), \c{~}
1786 computes the \i{one's complement} of its operand, \c{!} is the
1787 \i{logical negation} operator, and \c{SEG} provides the \i{segment address}
1788 of its operand (explained in more detail in \k{segwrt}).
1791 \H{segwrt} \i\c{SEG} and \i\c{WRT}
1793 When writing large 16-bit programs, which must be split into
1794 multiple \i{segments}, it is often necessary to be able to refer to
1795 the \I{segment address}segment part of the address of a symbol. NASM
1796 supports the \c{SEG} operator to perform this function.
1798 The \c{SEG} operator returns the \i\e{preferred} segment base of a
1799 symbol, defined as the segment base relative to which the offset of
1800 the symbol makes sense. So the code
1802 \c mov ax,seg symbol
1806 will load \c{ES:BX} with a valid pointer to the symbol \c{symbol}.
1808 Things can be more complex than this: since 16-bit segments and
1809 \i{groups} may \I{overlapping segments}overlap, you might occasionally
1810 want to refer to some symbol using a different segment base from the
1811 preferred one. NASM lets you do this, by the use of the \c{WRT}
1812 (With Reference To) keyword. So you can do things like
1814 \c mov ax,weird_seg ; weird_seg is a segment base
1816 \c mov bx,symbol wrt weird_seg
1818 to load \c{ES:BX} with a different, but functionally equivalent,
1819 pointer to the symbol \c{symbol}.
1821 NASM supports far (inter-segment) calls and jumps by means of the
1822 syntax \c{call segment:offset}, where \c{segment} and \c{offset}
1823 both represent immediate values. So to call a far procedure, you
1824 could code either of
1826 \c call (seg procedure):procedure
1827 \c call weird_seg:(procedure wrt weird_seg)
1829 (The parentheses are included for clarity, to show the intended
1830 parsing of the above instructions. They are not necessary in
1833 NASM supports the syntax \I\c{CALL FAR}\c{call far procedure} as a
1834 synonym for the first of the above usages. \c{JMP} works identically
1835 to \c{CALL} in these examples.
1837 To declare a \i{far pointer} to a data item in a data segment, you
1840 \c dw symbol, seg symbol
1842 NASM supports no convenient synonym for this, though you can always
1843 invent one using the macro processor.
1846 \H{strict} \i\c{STRICT}: Inhibiting Optimization
1848 When assembling with the optimizer set to level 2 or higher (see
1849 \k{opt-O}), NASM will use size specifiers (\c{BYTE}, \c{WORD},
1850 \c{DWORD}, \c{QWORD}, \c{TWORD}, \c{OWORD} or \c{YWORD}), but will
1851 give them the smallest possible size. The keyword \c{STRICT} can be
1852 used to inhibit optimization and force a particular operand to be
1853 emitted in the specified size. For example, with the optimizer on, and
1854 in \c{BITS 16} mode,
1858 is encoded in three bytes \c{66 6A 21}, whereas
1860 \c push strict dword 33
1862 is encoded in six bytes, with a full dword immediate operand \c{66 68
1865 With the optimizer off, the same code (six bytes) is generated whether
1866 the \c{STRICT} keyword was used or not.
1869 \H{crit} \i{Critical Expressions}
1871 Although NASM has an optional multi-pass optimizer, there are some
1872 expressions which must be resolvable on the first pass. These are
1873 called \e{Critical Expressions}.
1875 The first pass is used to determine the size of all the assembled
1876 code and data, so that the second pass, when generating all the
1877 code, knows all the symbol addresses the code refers to. So one
1878 thing NASM can't handle is code whose size depends on the value of a
1879 symbol declared after the code in question. For example,
1881 \c times (label-$) db 0
1882 \c label: db 'Where am I?'
1884 The argument to \i\c{TIMES} in this case could equally legally
1885 evaluate to anything at all; NASM will reject this example because
1886 it cannot tell the size of the \c{TIMES} line when it first sees it.
1887 It will just as firmly reject the slightly \I{paradox}paradoxical
1890 \c times (label-$+1) db 0
1891 \c label: db 'NOW where am I?'
1893 in which \e{any} value for the \c{TIMES} argument is by definition
1896 NASM rejects these examples by means of a concept called a
1897 \e{critical expression}, which is defined to be an expression whose
1898 value is required to be computable in the first pass, and which must
1899 therefore depend only on symbols defined before it. The argument to
1900 the \c{TIMES} prefix is a critical expression.
1902 \H{locallab} \i{Local Labels}
1904 NASM gives special treatment to symbols beginning with a \i{period}.
1905 A label beginning with a single period is treated as a \e{local}
1906 label, which means that it is associated with the previous non-local
1907 label. So, for example:
1909 \c label1 ; some code
1917 \c label2 ; some code
1925 In the above code fragment, each \c{JNE} instruction jumps to the
1926 line immediately before it, because the two definitions of \c{.loop}
1927 are kept separate by virtue of each being associated with the
1928 previous non-local label.
1930 This form of local label handling is borrowed from the old Amiga
1931 assembler \i{DevPac}; however, NASM goes one step further, in
1932 allowing access to local labels from other parts of the code. This
1933 is achieved by means of \e{defining} a local label in terms of the
1934 previous non-local label: the first definition of \c{.loop} above is
1935 really defining a symbol called \c{label1.loop}, and the second
1936 defines a symbol called \c{label2.loop}. So, if you really needed
1939 \c label3 ; some more code
1944 Sometimes it is useful - in a macro, for instance - to be able to
1945 define a label which can be referenced from anywhere but which
1946 doesn't interfere with the normal local-label mechanism. Such a
1947 label can't be non-local because it would interfere with subsequent
1948 definitions of, and references to, local labels; and it can't be
1949 local because the macro that defined it wouldn't know the label's
1950 full name. NASM therefore introduces a third type of label, which is
1951 probably only useful in macro definitions: if a label begins with
1952 the \I{label prefix}special prefix \i\c{..@}, then it does nothing
1953 to the local label mechanism. So you could code
1955 \c label1: ; a non-local label
1956 \c .local: ; this is really label1.local
1957 \c ..@foo: ; this is a special symbol
1958 \c label2: ; another non-local label
1959 \c .local: ; this is really label2.local
1961 \c jmp ..@foo ; this will jump three lines up
1963 NASM has the capacity to define other special symbols beginning with
1964 a double period: for example, \c{..start} is used to specify the
1965 entry point in the \c{obj} output format (see \k{dotdotstart}),
1966 \c{..imagebase} is used to find out the offset from a base address
1967 of the current image in the \c{win64} output format (see \k{win64pic}).
1968 So just keep in mind that symbols beginning with a double period are
1972 \C{preproc} The NASM \i{Preprocessor}
1974 NASM contains a powerful \i{macro processor}, which supports
1975 conditional assembly, multi-level file inclusion, two forms of macro
1976 (single-line and multi-line), and a `context stack' mechanism for
1977 extra macro power. Preprocessor directives all begin with a \c{%}
1980 The preprocessor collapses all lines which end with a backslash (\\)
1981 character into a single line. Thus:
1983 \c %define THIS_VERY_LONG_MACRO_NAME_IS_DEFINED_TO \\
1986 will work like a single-line macro without the backslash-newline
1989 \H{slmacro} \i{Single-Line Macros}
1991 \S{define} The Normal Way: \I\c{%idefine}\i\c{%define}
1993 Single-line macros are defined using the \c{%define} preprocessor
1994 directive. The definitions work in a similar way to C; so you can do
1997 \c %define ctrl 0x1F &
1998 \c %define param(a,b) ((a)+(a)*(b))
2000 \c mov byte [param(2,ebx)], ctrl 'D'
2002 which will expand to
2004 \c mov byte [(2)+(2)*(ebx)], 0x1F & 'D'
2006 When the expansion of a single-line macro contains tokens which
2007 invoke another macro, the expansion is performed at invocation time,
2008 not at definition time. Thus the code
2010 \c %define a(x) 1+b(x)
2015 will evaluate in the expected way to \c{mov ax,1+2*8}, even though
2016 the macro \c{b} wasn't defined at the time of definition of \c{a}.
2018 Macros defined with \c{%define} are \i{case sensitive}: after
2019 \c{%define foo bar}, only \c{foo} will expand to \c{bar}: \c{Foo} or
2020 \c{FOO} will not. By using \c{%idefine} instead of \c{%define} (the
2021 `i' stands for `insensitive') you can define all the case variants
2022 of a macro at once, so that \c{%idefine foo bar} would cause
2023 \c{foo}, \c{Foo}, \c{FOO}, \c{fOO} and so on all to expand to
2026 There is a mechanism which detects when a macro call has occurred as
2027 a result of a previous expansion of the same macro, to guard against
2028 \i{circular references} and infinite loops. If this happens, the
2029 preprocessor will only expand the first occurrence of the macro.
2032 \c %define a(x) 1+a(x)
2036 the macro \c{a(3)} will expand once, becoming \c{1+a(3)}, and will
2037 then expand no further. This behaviour can be useful: see \k{32c}
2038 for an example of its use.
2040 You can \I{overloading, single-line macros}overload single-line
2041 macros: if you write
2043 \c %define foo(x) 1+x
2044 \c %define foo(x,y) 1+x*y
2046 the preprocessor will be able to handle both types of macro call,
2047 by counting the parameters you pass; so \c{foo(3)} will become
2048 \c{1+3} whereas \c{foo(ebx,2)} will become \c{1+ebx*2}. However, if
2053 then no other definition of \c{foo} will be accepted: a macro with
2054 no parameters prohibits the definition of the same name as a macro
2055 \e{with} parameters, and vice versa.
2057 This doesn't prevent single-line macros being \e{redefined}: you can
2058 perfectly well define a macro with
2062 and then re-define it later in the same source file with
2066 Then everywhere the macro \c{foo} is invoked, it will be expanded
2067 according to the most recent definition. This is particularly useful
2068 when defining single-line macros with \c{%assign} (see \k{assign}).
2070 You can \i{pre-define} single-line macros using the `-d' option on
2071 the NASM command line: see \k{opt-d}.
2074 \S{xdefine} Resolving \c{%define}: \I\c{%ixdefine}\i\c{%xdefine}
2076 To have a reference to an embedded single-line macro resolved at the
2077 time that the embedding macro is \e{defined}, as opposed to when the
2078 embedding macro is \e{expanded}, you need a different mechanism to the
2079 one offered by \c{%define}. The solution is to use \c{%xdefine}, or
2080 it's \I{case sensitive}case-insensitive counterpart \c{%ixdefine}.
2082 Suppose you have the following code:
2085 \c %define isFalse isTrue
2094 In this case, \c{val1} is equal to 0, and \c{val2} is equal to 1.
2095 This is because, when a single-line macro is defined using
2096 \c{%define}, it is expanded only when it is called. As \c{isFalse}
2097 expands to \c{isTrue}, the expansion will be the current value of
2098 \c{isTrue}. The first time it is called that is 0, and the second
2101 If you wanted \c{isFalse} to expand to the value assigned to the
2102 embedded macro \c{isTrue} at the time that \c{isFalse} was defined,
2103 you need to change the above code to use \c{%xdefine}.
2105 \c %xdefine isTrue 1
2106 \c %xdefine isFalse isTrue
2107 \c %xdefine isTrue 0
2111 \c %xdefine isTrue 1
2115 Now, each time that \c{isFalse} is called, it expands to 1,
2116 as that is what the embedded macro \c{isTrue} expanded to at
2117 the time that \c{isFalse} was defined.
2120 \S{indmacro} \i{Macro Indirection}: \I\c{%[}\c{%[...]}
2122 The \c{%[...]} construct can be used to expand macros in contexts
2123 where macro expansion would otherwise not occur, including in the
2124 names other macros. For example, if you have a set of macros named
2125 \c{Foo16}, \c{Foo32} and \c{Foo64}, you could write:
2127 \c mov ax,Foo%[__BITS__] ; The Foo value
2129 to use the builtin macro \c{__BITS__} (see \k{bitsm}) to automatically
2130 select between them. Similarly, the two statements:
2132 \c %xdefine Bar Quux ; Expands due to %xdefine
2133 \c %define Bar %[Quux] ; Expands due to %[...]
2135 have, in fact, exactly the same effect.
2137 \c{%[...]} concatenates to adjacent tokens in the same way that
2138 multi-line macro parameters do, see \k{concat} for details.
2141 \S{concat%+} Concatenating Single Line Macro Tokens: \i\c{%+}
2143 Individual tokens in single line macros can be concatenated, to produce
2144 longer tokens for later processing. This can be useful if there are
2145 several similar macros that perform similar functions.
2147 Please note that a space is required after \c{%+}, in order to
2148 disambiguate it from the syntax \c{%+1} used in multiline macros.
2150 As an example, consider the following:
2152 \c %define BDASTART 400h ; Start of BIOS data area
2154 \c struc tBIOSDA ; its structure
2160 Now, if we need to access the elements of tBIOSDA in different places,
2163 \c mov ax,BDASTART + tBIOSDA.COM1addr
2164 \c mov bx,BDASTART + tBIOSDA.COM2addr
2166 This will become pretty ugly (and tedious) if used in many places, and
2167 can be reduced in size significantly by using the following macro:
2169 \c ; Macro to access BIOS variables by their names (from tBDA):
2171 \c %define BDA(x) BDASTART + tBIOSDA. %+ x
2173 Now the above code can be written as:
2175 \c mov ax,BDA(COM1addr)
2176 \c mov bx,BDA(COM2addr)
2178 Using this feature, we can simplify references to a lot of macros (and,
2179 in turn, reduce typing errors).
2182 \S{selfref%?} The Macro Name Itself: \i\c{%?} and \i\c{%??}
2184 The special symbols \c{%?} and \c{%??} can be used to reference the
2185 macro name itself inside a macro expansion, this is supported for both
2186 single-and multi-line macros. \c{%?} refers to the macro name as
2187 \e{invoked}, whereas \c{%??} refers to the macro name as
2188 \e{declared}. The two are always the same for case-sensitive
2189 macros, but for case-insensitive macros, they can differ.
2193 \c %idefine Foo mov %?,%??
2205 \c %idefine keyword $%?
2207 can be used to make a keyword "disappear", for example in case a new
2208 instruction has been used as a label in older code. For example:
2210 \c %idefine pause $%? ; Hide the PAUSE instruction
2213 \S{undef} Undefining Single-Line Macros: \i\c{%undef}
2215 Single-line macros can be removed with the \c{%undef} directive. For
2216 example, the following sequence:
2223 will expand to the instruction \c{mov eax, foo}, since after
2224 \c{%undef} the macro \c{foo} is no longer defined.
2226 Macros that would otherwise be pre-defined can be undefined on the
2227 command-line using the `-u' option on the NASM command line: see
2231 \S{assign} \i{Preprocessor Variables}: \i\c{%assign}
2233 An alternative way to define single-line macros is by means of the
2234 \c{%assign} command (and its \I{case sensitive}case-insensitive
2235 counterpart \i\c{%iassign}, which differs from \c{%assign} in
2236 exactly the same way that \c{%idefine} differs from \c{%define}).
2238 \c{%assign} is used to define single-line macros which take no
2239 parameters and have a numeric value. This value can be specified in
2240 the form of an expression, and it will be evaluated once, when the
2241 \c{%assign} directive is processed.
2243 Like \c{%define}, macros defined using \c{%assign} can be re-defined
2244 later, so you can do things like
2248 to increment the numeric value of a macro.
2250 \c{%assign} is useful for controlling the termination of \c{%rep}
2251 preprocessor loops: see \k{rep} for an example of this. Another
2252 use for \c{%assign} is given in \k{16c} and \k{32c}.
2254 The expression passed to \c{%assign} is a \i{critical expression}
2255 (see \k{crit}), and must also evaluate to a pure number (rather than
2256 a relocatable reference such as a code or data address, or anything
2257 involving a register).
2260 \S{defstr} Defining Strings: \I\c{%idefstr}\i\c{%defstr}
2262 \c{%defstr}, and its case-insensitive counterpart \c{%idefstr}, define
2263 or redefine a single-line macro without parameters but converts the
2264 entire right-hand side, after macro expansion, to a quoted string
2269 \c %defstr test TEST
2273 \c %define test 'TEST'
2275 This can be used, for example, with the \c{%!} construct (see
2278 \c %defstr PATH %!PATH ; The operating system PATH variable
2281 \S{deftok} Defining Tokens: \I\c{%ideftok}\i\c{%deftok}
2283 \c{%deftok}, and its case-insensitive counterpart \c{%ideftok}, define
2284 or redefine a single-line macro without parameters but converts the
2285 second parameter, after string conversion, to a sequence of tokens.
2289 \c %deftok test 'TEST'
2293 \c %define test TEST
2296 \H{strlen} \i{String Manipulation in Macros}
2298 It's often useful to be able to handle strings in macros. NASM
2299 supports a few simple string handling macro operators from which
2300 more complex operations can be constructed.
2302 All the string operators define or redefine a value (either a string
2303 or a numeric value) to a single-line macro. When producing a string
2304 value, it may change the style of quoting of the input string or
2305 strings, and possibly use \c{\\}-escapes inside \c{`}-quoted strings.
2307 \S{strcat} \i{Concatenating Strings}: \i\c{%strcat}
2309 The \c{%strcat} operator concatenates quoted strings and assign them to
2310 a single-line macro.
2314 \c %strcat alpha "Alpha: ", '12" screen'
2316 ... would assign the value \c{'Alpha: 12" screen'} to \c{alpha}.
2319 \c %strcat beta '"foo"\', "'bar'"
2321 ... would assign the value \c{`"foo"\\\\'bar'`} to \c{beta}.
2323 The use of commas to separate strings is permitted but optional.
2326 \S{strlen} \i{String Length}: \i\c{%strlen}
2328 The \c{%strlen} operator assigns the length of a string to a macro.
2331 \c %strlen charcnt 'my string'
2333 In this example, \c{charcnt} would receive the value 9, just as
2334 if an \c{%assign} had been used. In this example, \c{'my string'}
2335 was a literal string but it could also have been a single-line
2336 macro that expands to a string, as in the following example:
2338 \c %define sometext 'my string'
2339 \c %strlen charcnt sometext
2341 As in the first case, this would result in \c{charcnt} being
2342 assigned the value of 9.
2345 \S{substr} \i{Extracting Substrings}: \i\c{%substr}
2347 Individual letters or substrings in strings can be extracted using the
2348 \c{%substr} operator. An example of its use is probably more useful
2349 than the description:
2351 \c %substr mychar 'xyzw' 1 ; equivalent to %define mychar 'x'
2352 \c %substr mychar 'xyzw' 2 ; equivalent to %define mychar 'y'
2353 \c %substr mychar 'xyzw' 3 ; equivalent to %define mychar 'z'
2354 \c %substr mychar 'xyzw' 2,2 ; equivalent to %define mychar 'yz'
2355 \c %substr mychar 'xyzw' 2,-1 ; equivalent to %define mychar 'yzw'
2356 \c %substr mychar 'xyzw' 2,-2 ; equivalent to %define mychar 'yz'
2358 As with \c{%strlen} (see \k{strlen}), the first parameter is the
2359 single-line macro to be created and the second is the string. The
2360 third parameter specifies the first character to be selected, and the
2361 optional fourth parameter preceeded by comma) is the length. Note
2362 that the first index is 1, not 0 and the last index is equal to the
2363 value that \c{%strlen} would assign given the same string. Index
2364 values out of range result in an empty string. A negative length
2365 means "until N-1 characters before the end of string", i.e. \c{-1}
2366 means until end of string, \c{-2} until one character before, etc.
2369 \H{mlmacro} \i{Multi-Line Macros}: \I\c{%imacro}\i\c{%macro}
2371 Multi-line macros are much more like the type of macro seen in MASM
2372 and TASM: a multi-line macro definition in NASM looks something like
2375 \c %macro prologue 1
2383 This defines a C-like function prologue as a macro: so you would
2384 invoke the macro with a call such as
2386 \c myfunc: prologue 12
2388 which would expand to the three lines of code
2394 The number \c{1} after the macro name in the \c{%macro} line defines
2395 the number of parameters the macro \c{prologue} expects to receive.
2396 The use of \c{%1} inside the macro definition refers to the first
2397 parameter to the macro call. With a macro taking more than one
2398 parameter, subsequent parameters would be referred to as \c{%2},
2401 Multi-line macros, like single-line macros, are \i{case-sensitive},
2402 unless you define them using the alternative directive \c{%imacro}.
2404 If you need to pass a comma as \e{part} of a parameter to a
2405 multi-line macro, you can do that by enclosing the entire parameter
2406 in \I{braces, around macro parameters}braces. So you could code
2415 \c silly 'a', letter_a ; letter_a: db 'a'
2416 \c silly 'ab', string_ab ; string_ab: db 'ab'
2417 \c silly {13,10}, crlf ; crlf: db 13,10
2420 \S{mlrmacro} \i{Recursive Multi-Line Macros}: \I\c{%irmacro}\i\c{%rmacro}
2422 A multi-line macro cannot be referenced within itself, in order to
2423 prevent accidental infinite recursion and allow instruction overloading.
2425 Recursive multi-line macros allow for self-referencing, with the
2426 caveat that the user is aware of the existence, use and purpose of
2427 recursive multi-line macros. There is also a generous, but sane, upper
2428 limit to the number of recursions, in order to prevent run-away memory
2429 consumption in case of accidental infinite recursion.
2431 As with non-recursive multi-line macros, recursive multi-line macros are
2432 \i{case-sensitive}, unless you define them using the alternative
2433 directive \c{%irmacro}.
2436 \S{mlmacover} Overloading Multi-Line Macros\I{overloading, multi-line macros}
2438 As with single-line macros, multi-line macros can be overloaded by
2439 defining the same macro name several times with different numbers of
2440 parameters. This time, no exception is made for macros with no
2441 parameters at all. So you could define
2443 \c %macro prologue 0
2450 to define an alternative form of the function prologue which
2451 allocates no local stack space.
2453 Sometimes, however, you might want to `overload' a machine
2454 instruction; for example, you might want to define
2463 so that you could code
2465 \c push ebx ; this line is not a macro call
2466 \c push eax,ecx ; but this one is
2468 Ordinarily, NASM will give a warning for the first of the above two
2469 lines, since \c{push} is now defined to be a macro, and is being
2470 invoked with a number of parameters for which no definition has been
2471 given. The correct code will still be generated, but the assembler
2472 will give a warning. This warning can be disabled by the use of the
2473 \c{-w-macro-params} command-line option (see \k{opt-w}).
2476 \S{maclocal} \i{Macro-Local Labels}
2478 NASM allows you to define labels within a multi-line macro
2479 definition in such a way as to make them local to the macro call: so
2480 calling the same macro multiple times will use a different label
2481 each time. You do this by prefixing \i\c{%%} to the label name. So
2482 you can invent an instruction which executes a \c{RET} if the \c{Z}
2483 flag is set by doing this:
2493 You can call this macro as many times as you want, and every time
2494 you call it NASM will make up a different `real' name to substitute
2495 for the label \c{%%skip}. The names NASM invents are of the form
2496 \c{..@2345.skip}, where the number 2345 changes with every macro
2497 call. The \i\c{..@} prefix prevents macro-local labels from
2498 interfering with the local label mechanism, as described in
2499 \k{locallab}. You should avoid defining your own labels in this form
2500 (the \c{..@} prefix, then a number, then another period) in case
2501 they interfere with macro-local labels.
2504 \S{mlmacgre} \i{Greedy Macro Parameters}
2506 Occasionally it is useful to define a macro which lumps its entire
2507 command line into one parameter definition, possibly after
2508 extracting one or two smaller parameters from the front. An example
2509 might be a macro to write a text string to a file in MS-DOS, where
2510 you might want to be able to write
2512 \c writefile [filehandle],"hello, world",13,10
2514 NASM allows you to define the last parameter of a macro to be
2515 \e{greedy}, meaning that if you invoke the macro with more
2516 parameters than it expects, all the spare parameters get lumped into
2517 the last defined one along with the separating commas. So if you
2520 \c %macro writefile 2+
2526 \c mov cx,%%endstr-%%str
2533 then the example call to \c{writefile} above will work as expected:
2534 the text before the first comma, \c{[filehandle]}, is used as the
2535 first macro parameter and expanded when \c{%1} is referred to, and
2536 all the subsequent text is lumped into \c{%2} and placed after the
2539 The greedy nature of the macro is indicated to NASM by the use of
2540 the \I{+ modifier}\c{+} sign after the parameter count on the
2543 If you define a greedy macro, you are effectively telling NASM how
2544 it should expand the macro given \e{any} number of parameters from
2545 the actual number specified up to infinity; in this case, for
2546 example, NASM now knows what to do when it sees a call to
2547 \c{writefile} with 2, 3, 4 or more parameters. NASM will take this
2548 into account when overloading macros, and will not allow you to
2549 define another form of \c{writefile} taking 4 parameters (for
2552 Of course, the above macro could have been implemented as a
2553 non-greedy macro, in which case the call to it would have had to
2556 \c writefile [filehandle], {"hello, world",13,10}
2558 NASM provides both mechanisms for putting \i{commas in macro
2559 parameters}, and you choose which one you prefer for each macro
2562 See \k{sectmac} for a better way to write the above macro.
2564 \S{mlmacrange} \i{Macro Parameters Range}
2566 NASM allows you to expand parameters via special construction \c{%\{x:y\}}
2567 where \c{x} is the first parameter index and \c{y} is the last. Any index can
2568 be either negative or positive but must never be zero.
2578 expands to \c{3,4,5} range.
2580 Even more, the parameters can be reversed so that
2588 expands to \c{5,4,3} range.
2590 But even this is not the last. The parameters can be addressed via negative
2591 indices so NASM will count them reversed. The ones who know Python may see
2600 expands to \c{6,5,4} range.
2602 Note that NASM uses \i{comma} to separate parameters being expanded.
2604 By the way, here is a trick - you might use the index \c{%{-1:-1}}
2605 which gives you the \i{last} argument passed to a macro.
2607 \S{mlmacdef} \i{Default Macro Parameters}
2609 NASM also allows you to define a multi-line macro with a \e{range}
2610 of allowable parameter counts. If you do this, you can specify
2611 defaults for \i{omitted parameters}. So, for example:
2613 \c %macro die 0-1 "Painful program death has occurred."
2621 This macro (which makes use of the \c{writefile} macro defined in
2622 \k{mlmacgre}) can be called with an explicit error message, which it
2623 will display on the error output stream before exiting, or it can be
2624 called with no parameters, in which case it will use the default
2625 error message supplied in the macro definition.
2627 In general, you supply a minimum and maximum number of parameters
2628 for a macro of this type; the minimum number of parameters are then
2629 required in the macro call, and then you provide defaults for the
2630 optional ones. So if a macro definition began with the line
2632 \c %macro foobar 1-3 eax,[ebx+2]
2634 then it could be called with between one and three parameters, and
2635 \c{%1} would always be taken from the macro call. \c{%2}, if not
2636 specified by the macro call, would default to \c{eax}, and \c{%3} if
2637 not specified would default to \c{[ebx+2]}.
2639 You can provide extra information to a macro by providing
2640 too many default parameters:
2642 \c %macro quux 1 something
2644 This will trigger a warning by default; see \k{opt-w} for
2646 When \c{quux} is invoked, it receives not one but two parameters.
2647 \c{something} can be referred to as \c{%2}. The difference
2648 between passing \c{something} this way and writing \c{something}
2649 in the macro body is that with this way \c{something} is evaluated
2650 when the macro is defined, not when it is expanded.
2652 You may omit parameter defaults from the macro definition, in which
2653 case the parameter default is taken to be blank. This can be useful
2654 for macros which can take a variable number of parameters, since the
2655 \i\c{%0} token (see \k{percent0}) allows you to determine how many
2656 parameters were really passed to the macro call.
2658 This defaulting mechanism can be combined with the greedy-parameter
2659 mechanism; so the \c{die} macro above could be made more powerful,
2660 and more useful, by changing the first line of the definition to
2662 \c %macro die 0-1+ "Painful program death has occurred.",13,10
2664 The maximum parameter count can be infinite, denoted by \c{*}. In
2665 this case, of course, it is impossible to provide a \e{full} set of
2666 default parameters. Examples of this usage are shown in \k{rotate}.
2669 \S{percent0} \i\c{%0}: \I{counting macro parameters}Macro Parameter Counter
2671 The parameter reference \c{%0} will return a numeric constant giving the
2672 number of parameters received, that is, if \c{%0} is n then \c{%}n is the
2673 last parameter. \c{%0} is mostly useful for macros that can take a variable
2674 number of parameters. It can be used as an argument to \c{%rep}
2675 (see \k{rep}) in order to iterate through all the parameters of a macro.
2676 Examples are given in \k{rotate}.
2679 \S{percent00} \i\c{%00}: \I{label preceeding macro}Label Preceeding Macro
2681 \c{%00} will return the label preceeding the macro invocation, if any. The
2682 label must be on the same line as the macro invocation, may be a local label
2683 (see \k{locallab}), and need not end in a colon.
2686 \S{rotate} \i\c{%rotate}: \i{Rotating Macro Parameters}
2688 Unix shell programmers will be familiar with the \I{shift
2689 command}\c{shift} shell command, which allows the arguments passed
2690 to a shell script (referenced as \c{$1}, \c{$2} and so on) to be
2691 moved left by one place, so that the argument previously referenced
2692 as \c{$2} becomes available as \c{$1}, and the argument previously
2693 referenced as \c{$1} is no longer available at all.
2695 NASM provides a similar mechanism, in the form of \c{%rotate}. As
2696 its name suggests, it differs from the Unix \c{shift} in that no
2697 parameters are lost: parameters rotated off the left end of the
2698 argument list reappear on the right, and vice versa.
2700 \c{%rotate} is invoked with a single numeric argument (which may be
2701 an expression). The macro parameters are rotated to the left by that
2702 many places. If the argument to \c{%rotate} is negative, the macro
2703 parameters are rotated to the right.
2705 \I{iterating over macro parameters}So a pair of macros to save and
2706 restore a set of registers might work as follows:
2708 \c %macro multipush 1-*
2717 This macro invokes the \c{PUSH} instruction on each of its arguments
2718 in turn, from left to right. It begins by pushing its first
2719 argument, \c{%1}, then invokes \c{%rotate} to move all the arguments
2720 one place to the left, so that the original second argument is now
2721 available as \c{%1}. Repeating this procedure as many times as there
2722 were arguments (achieved by supplying \c{%0} as the argument to
2723 \c{%rep}) causes each argument in turn to be pushed.
2725 Note also the use of \c{*} as the maximum parameter count,
2726 indicating that there is no upper limit on the number of parameters
2727 you may supply to the \i\c{multipush} macro.
2729 It would be convenient, when using this macro, to have a \c{POP}
2730 equivalent, which \e{didn't} require the arguments to be given in
2731 reverse order. Ideally, you would write the \c{multipush} macro
2732 call, then cut-and-paste the line to where the pop needed to be
2733 done, and change the name of the called macro to \c{multipop}, and
2734 the macro would take care of popping the registers in the opposite
2735 order from the one in which they were pushed.
2737 This can be done by the following definition:
2739 \c %macro multipop 1-*
2748 This macro begins by rotating its arguments one place to the
2749 \e{right}, so that the original \e{last} argument appears as \c{%1}.
2750 This is then popped, and the arguments are rotated right again, so
2751 the second-to-last argument becomes \c{%1}. Thus the arguments are
2752 iterated through in reverse order.
2755 \S{concat} \i{Concatenating Macro Parameters}
2757 NASM can concatenate macro parameters and macro indirection constructs
2758 on to other text surrounding them. This allows you to declare a family
2759 of symbols, for example, in a macro definition. If, for example, you
2760 wanted to generate a table of key codes along with offsets into the
2761 table, you could code something like
2763 \c %macro keytab_entry 2
2765 \c keypos%1 equ $-keytab
2771 \c keytab_entry F1,128+1
2772 \c keytab_entry F2,128+2
2773 \c keytab_entry Return,13
2775 which would expand to
2778 \c keyposF1 equ $-keytab
2780 \c keyposF2 equ $-keytab
2782 \c keyposReturn equ $-keytab
2785 You can just as easily concatenate text on to the other end of a
2786 macro parameter, by writing \c{%1foo}.
2788 If you need to append a \e{digit} to a macro parameter, for example
2789 defining labels \c{foo1} and \c{foo2} when passed the parameter
2790 \c{foo}, you can't code \c{%11} because that would be taken as the
2791 eleventh macro parameter. Instead, you must code
2792 \I{braces, after % sign}\c{%\{1\}1}, which will separate the first
2793 \c{1} (giving the number of the macro parameter) from the second
2794 (literal text to be concatenated to the parameter).
2796 This concatenation can also be applied to other preprocessor in-line
2797 objects, such as macro-local labels (\k{maclocal}) and context-local
2798 labels (\k{ctxlocal}). In all cases, ambiguities in syntax can be
2799 resolved by enclosing everything after the \c{%} sign and before the
2800 literal text in braces: so \c{%\{%foo\}bar} concatenates the text
2801 \c{bar} to the end of the real name of the macro-local label
2802 \c{%%foo}. (This is unnecessary, since the form NASM uses for the
2803 real names of macro-local labels means that the two usages
2804 \c{%\{%foo\}bar} and \c{%%foobar} would both expand to the same
2805 thing anyway; nevertheless, the capability is there.)
2807 The single-line macro indirection construct, \c{%[...]}
2808 (\k{indmacro}), behaves the same way as macro parameters for the
2809 purpose of concatenation.
2811 See also the \c{%+} operator, \k{concat%+}.
2814 \S{mlmaccc} \i{Condition Codes as Macro Parameters}
2816 NASM can give special treatment to a macro parameter which contains
2817 a condition code. For a start, you can refer to the macro parameter
2818 \c{%1} by means of the alternative syntax \i\c{%+1}, which informs
2819 NASM that this macro parameter is supposed to contain a condition
2820 code, and will cause the preprocessor to report an error message if
2821 the macro is called with a parameter which is \e{not} a valid
2824 Far more usefully, though, you can refer to the macro parameter by
2825 means of \i\c{%-1}, which NASM will expand as the \e{inverse}
2826 condition code. So the \c{retz} macro defined in \k{maclocal} can be
2827 replaced by a general \i{conditional-return macro} like this:
2837 This macro can now be invoked using calls like \c{retc ne}, which
2838 will cause the conditional-jump instruction in the macro expansion
2839 to come out as \c{JE}, or \c{retc po} which will make the jump a
2842 The \c{%+1} macro-parameter reference is quite happy to interpret
2843 the arguments \c{CXZ} and \c{ECXZ} as valid condition codes;
2844 however, \c{%-1} will report an error if passed either of these,
2845 because no inverse condition code exists.
2848 \S{nolist} \i{Disabling Listing Expansion}\I\c{.nolist}
2850 When NASM is generating a listing file from your program, it will
2851 generally expand multi-line macros by means of writing the macro
2852 call and then listing each line of the expansion. This allows you to
2853 see which instructions in the macro expansion are generating what
2854 code; however, for some macros this clutters the listing up
2857 NASM therefore provides the \c{.nolist} qualifier, which you can
2858 include in a macro definition to inhibit the expansion of the macro
2859 in the listing file. The \c{.nolist} qualifier comes directly after
2860 the number of parameters, like this:
2862 \c %macro foo 1.nolist
2866 \c %macro bar 1-5+.nolist a,b,c,d,e,f,g,h
2868 \S{unmacro} Undefining Multi-Line Macros: \i\c{%unmacro}
2870 Multi-line macros can be removed with the \c{%unmacro} directive.
2871 Unlike the \c{%undef} directive, however, \c{%unmacro} takes an
2872 argument specification, and will only remove \i{exact matches} with
2873 that argument specification.
2882 removes the previously defined macro \c{foo}, but
2889 does \e{not} remove the macro \c{bar}, since the argument
2890 specification does not match exactly.
2893 \S{exitmacro} Exiting Multi-Line Macros: \i\c{%exitmacro}
2895 Multi-line macro expansions can be arbitrarily terminated with
2896 the \c{%exitmacro} directive.
2909 \H{condasm} \i{Conditional Assembly}\I\c{%if}
2911 Similarly to the C preprocessor, NASM allows sections of a source
2912 file to be assembled only if certain conditions are met. The general
2913 syntax of this feature looks like this:
2916 \c ; some code which only appears if <condition> is met
2917 \c %elif<condition2>
2918 \c ; only appears if <condition> is not met but <condition2> is
2920 \c ; this appears if neither <condition> nor <condition2> was met
2923 The inverse forms \i\c{%ifn} and \i\c{%elifn} are also supported.
2925 The \i\c{%else} clause is optional, as is the \i\c{%elif} clause.
2926 You can have more than one \c{%elif} clause as well.
2928 There are a number of variants of the \c{%if} directive. Each has its
2929 corresponding \c{%elif}, \c{%ifn}, and \c{%elifn} directives; for
2930 example, the equivalents to the \c{%ifdef} directive are \c{%elifdef},
2931 \c{%ifndef}, and \c{%elifndef}.
2933 \S{ifdef} \i\c{%ifdef}: Testing Single-Line Macro Existence\I{testing,
2934 single-line macro existence}
2936 Beginning a conditional-assembly block with the line \c{%ifdef
2937 MACRO} will assemble the subsequent code if, and only if, a
2938 single-line macro called \c{MACRO} is defined. If not, then the
2939 \c{%elif} and \c{%else} blocks (if any) will be processed instead.
2941 For example, when debugging a program, you might want to write code
2944 \c ; perform some function
2946 \c writefile 2,"Function performed successfully",13,10
2948 \c ; go and do something else
2950 Then you could use the command-line option \c{-dDEBUG} to create a
2951 version of the program which produced debugging messages, and remove
2952 the option to generate the final release version of the program.
2954 You can test for a macro \e{not} being defined by using
2955 \i\c{%ifndef} instead of \c{%ifdef}. You can also test for macro
2956 definitions in \c{%elif} blocks by using \i\c{%elifdef} and
2960 \S{ifmacro} \i\c{%ifmacro}: Testing Multi-Line Macro
2961 Existence\I{testing, multi-line macro existence}
2963 The \c{%ifmacro} directive operates in the same way as the \c{%ifdef}
2964 directive, except that it checks for the existence of a multi-line macro.
2966 For example, you may be working with a large project and not have control
2967 over the macros in a library. You may want to create a macro with one
2968 name if it doesn't already exist, and another name if one with that name
2971 The \c{%ifmacro} is considered true if defining a macro with the given name
2972 and number of arguments would cause a definitions conflict. For example:
2974 \c %ifmacro MyMacro 1-3
2976 \c %error "MyMacro 1-3" causes a conflict with an existing macro.
2980 \c %macro MyMacro 1-3
2982 \c ; insert code to define the macro
2988 This will create the macro "MyMacro 1-3" if no macro already exists which
2989 would conflict with it, and emits a warning if there would be a definition
2992 You can test for the macro not existing by using the \i\c{%ifnmacro} instead
2993 of \c{%ifmacro}. Additional tests can be performed in \c{%elif} blocks by using
2994 \i\c{%elifmacro} and \i\c{%elifnmacro}.
2997 \S{ifctx} \i\c{%ifctx}: Testing the Context Stack\I{testing, context
3000 The conditional-assembly construct \c{%ifctx} will cause the
3001 subsequent code to be assembled if and only if the top context on
3002 the preprocessor's context stack has the same name as one of the arguments.
3003 As with \c{%ifdef}, the inverse and \c{%elif} forms \i\c{%ifnctx},
3004 \i\c{%elifctx} and \i\c{%elifnctx} are also supported.
3006 For more details of the context stack, see \k{ctxstack}. For a
3007 sample use of \c{%ifctx}, see \k{blockif}.
3010 \S{if} \i\c{%if}: Testing Arbitrary Numeric Expressions\I{testing,
3011 arbitrary numeric expressions}
3013 The conditional-assembly construct \c{%if expr} will cause the
3014 subsequent code to be assembled if and only if the value of the
3015 numeric expression \c{expr} is non-zero. An example of the use of
3016 this feature is in deciding when to break out of a \c{%rep}
3017 preprocessor loop: see \k{rep} for a detailed example.
3019 The expression given to \c{%if}, and its counterpart \i\c{%elif}, is
3020 a critical expression (see \k{crit}).
3022 \c{%if} extends the normal NASM expression syntax, by providing a
3023 set of \i{relational operators} which are not normally available in
3024 expressions. The operators \i\c{=}, \i\c{<}, \i\c{>}, \i\c{<=},
3025 \i\c{>=} and \i\c{<>} test equality, less-than, greater-than,
3026 less-or-equal, greater-or-equal and not-equal respectively. The
3027 C-like forms \i\c{==} and \i\c{!=} are supported as alternative
3028 forms of \c{=} and \c{<>}. In addition, low-priority logical
3029 operators \i\c{&&}, \i\c{^^} and \i\c{||} are provided, supplying
3030 \i{logical AND}, \i{logical XOR} and \i{logical OR}. These work like
3031 the C logical operators (although C has no logical XOR), in that
3032 they always return either 0 or 1, and treat any non-zero input as 1
3033 (so that \c{^^}, for example, returns 1 if exactly one of its inputs
3034 is zero, and 0 otherwise). The relational operators also return 1
3035 for true and 0 for false.
3037 Like other \c{%if} constructs, \c{%if} has a counterpart
3038 \i\c{%elif}, and negative forms \i\c{%ifn} and \i\c{%elifn}.
3040 \S{ifidn} \i\c{%ifidn} and \i\c{%ifidni}: Testing Exact Text
3041 Identity\I{testing, exact text identity}
3043 The construct \c{%ifidn text1,text2} will cause the subsequent code
3044 to be assembled if and only if \c{text1} and \c{text2}, after
3045 expanding single-line macros, are identical pieces of text.
3046 Differences in white space are not counted.
3048 \c{%ifidni} is similar to \c{%ifidn}, but is \i{case-insensitive}.
3050 For example, the following macro pushes a register or number on the
3051 stack, and allows you to treat \c{IP} as a real register:
3053 \c %macro pushparam 1
3064 Like other \c{%if} constructs, \c{%ifidn} has a counterpart
3065 \i\c{%elifidn}, and negative forms \i\c{%ifnidn} and \i\c{%elifnidn}.
3066 Similarly, \c{%ifidni} has counterparts \i\c{%elifidni},
3067 \i\c{%ifnidni} and \i\c{%elifnidni}.
3069 \S{iftyp} \i\c{%ifid}, \i\c{%ifnum}, \i\c{%ifstr}: Testing Token
3070 Types\I{testing, token types}
3072 Some macros will want to perform different tasks depending on
3073 whether they are passed a number, a string, or an identifier. For
3074 example, a string output macro might want to be able to cope with
3075 being passed either a string constant or a pointer to an existing
3078 The conditional assembly construct \c{%ifid}, taking one parameter
3079 (which may be blank), assembles the subsequent code if and only if
3080 the first token in the parameter exists and is an identifier.
3081 \c{%ifnum} works similarly, but tests for the token being a numeric
3082 constant; \c{%ifstr} tests for it being a string.
3084 For example, the \c{writefile} macro defined in \k{mlmacgre} can be
3085 extended to take advantage of \c{%ifstr} in the following fashion:
3087 \c %macro writefile 2-3+
3096 \c %%endstr: mov dx,%%str
3097 \c mov cx,%%endstr-%%str
3108 Then the \c{writefile} macro can cope with being called in either of
3109 the following two ways:
3111 \c writefile [file], strpointer, length
3112 \c writefile [file], "hello", 13, 10
3114 In the first, \c{strpointer} is used as the address of an
3115 already-declared string, and \c{length} is used as its length; in
3116 the second, a string is given to the macro, which therefore declares
3117 it itself and works out the address and length for itself.
3119 Note the use of \c{%if} inside the \c{%ifstr}: this is to detect
3120 whether the macro was passed two arguments (so the string would be a
3121 single string constant, and \c{db %2} would be adequate) or more (in
3122 which case, all but the first two would be lumped together into
3123 \c{%3}, and \c{db %2,%3} would be required).
3125 The usual \I\c{%elifid}\I\c{%elifnum}\I\c{%elifstr}\c{%elif}...,
3126 \I\c{%ifnid}\I\c{%ifnnum}\I\c{%ifnstr}\c{%ifn}..., and
3127 \I\c{%elifnid}\I\c{%elifnnum}\I\c{%elifnstr}\c{%elifn}... versions
3128 exist for each of \c{%ifid}, \c{%ifnum} and \c{%ifstr}.
3130 \S{iftoken} \i\c{%iftoken}: Test for a Single Token
3132 Some macros will want to do different things depending on if it is
3133 passed a single token (e.g. paste it to something else using \c{%+})
3134 versus a multi-token sequence.
3136 The conditional assembly construct \c{%iftoken} assembles the
3137 subsequent code if and only if the expanded parameters consist of
3138 exactly one token, possibly surrounded by whitespace.
3144 will assemble the subsequent code, but
3148 will not, since \c{-1} contains two tokens: the unary minus operator
3149 \c{-}, and the number \c{1}.
3151 The usual \i\c{%eliftoken}, \i\c\{%ifntoken}, and \i\c{%elifntoken}
3152 variants are also provided.
3154 \S{ifempty} \i\c{%ifempty}: Test for Empty Expansion
3156 The conditional assembly construct \c{%ifempty} assembles the
3157 subsequent code if and only if the expanded parameters do not contain
3158 any tokens at all, whitespace excepted.
3160 The usual \i\c{%elifempty}, \i\c\{%ifnempty}, and \i\c{%elifnempty}
3161 variants are also provided.
3163 \S{ifenv} \i\c{%ifenv}: Test If Environment Variable Exists
3165 The conditional assembly construct \c{%ifenv} assembles the
3166 subsequent code if and only if the environment variable referenced by
3167 the \c{%!<env>} directive exists.
3169 The usual \i\c{%elifenv}, \i\c\{%ifnenv}, and \i\c{%elifnenv}
3170 variants are also provided.
3172 Just as for \c{%!<env>} the argument should be written as a string if
3173 it contains characters that would not be legal in an identifier. See
3176 \H{rep} \i{Preprocessor Loops}\I{repeating code}: \i\c{%rep}
3178 NASM's \c{TIMES} prefix, though useful, cannot be used to invoke a
3179 multi-line macro multiple times, because it is processed by NASM
3180 after macros have already been expanded. Therefore NASM provides
3181 another form of loop, this time at the preprocessor level: \c{%rep}.
3183 The directives \c{%rep} and \i\c{%endrep} (\c{%rep} takes a numeric
3184 argument, which can be an expression; \c{%endrep} takes no
3185 arguments) can be used to enclose a chunk of code, which is then
3186 replicated as many times as specified by the preprocessor:
3190 \c inc word [table+2*i]
3194 This will generate a sequence of 64 \c{INC} instructions,
3195 incrementing every word of memory from \c{[table]} to
3198 For more complex termination conditions, or to break out of a repeat
3199 loop part way along, you can use the \i\c{%exitrep} directive to
3200 terminate the loop, like this:
3215 \c fib_number equ ($-fibonacci)/2
3217 This produces a list of all the Fibonacci numbers that will fit in
3218 16 bits. Note that a maximum repeat count must still be given to
3219 \c{%rep}. This is to prevent the possibility of NASM getting into an
3220 infinite loop in the preprocessor, which (on multitasking or
3221 multi-user systems) would typically cause all the system memory to
3222 be gradually used up and other applications to start crashing.
3224 Note a maximum repeat count is limited by 62 bit number, though it
3225 is hardly possible that you ever need anything bigger.
3228 \H{while} \i{Conditional Loops}: \i\c{%while}
3230 The directives \c{%while} and \i\c{%endwhile} combine preprocessor
3231 loops with conditional assembly, allowing the enclosed chunk of
3232 code to be replicated as long as certain conditions are met:
3234 \c %while<condition>
3235 \c ; some code which only repeats while <condition> is met
3238 \S{exitwhile} Exiting Conditional Loops: \i\c{%exitwhile}
3240 Conditional loops can be arbitrarily terminated with the
3241 \i\c{%exitwhile} directive.
3245 \c %while<condition>
3246 \c %if<some other condition>
3249 \c ; some code which only repeats while <condition> is met
3253 \H{files} Source Files and Dependencies
3255 These commands allow you to split your sources into multiple files.
3257 \S{include} \i\c{%include}: \i{Including Other Files}
3259 Using, once again, a very similar syntax to the C preprocessor,
3260 NASM's preprocessor lets you include other source files into your
3261 code. This is done by the use of the \i\c{%include} directive:
3263 \c %include "macros.mac"
3265 will include the contents of the file \c{macros.mac} into the source
3266 file containing the \c{%include} directive.
3268 Include files are \I{searching for include files}searched for in the
3269 current directory (the directory you're in when you run NASM, as
3270 opposed to the location of the NASM executable or the location of
3271 the source file), plus any directories specified on the NASM command
3272 line using the \c{-i} option.
3274 The standard C idiom for preventing a file being included more than
3275 once is just as applicable in NASM: if the file \c{macros.mac} has
3278 \c %ifndef MACROS_MAC
3279 \c %define MACROS_MAC
3280 \c ; now define some macros
3283 then including the file more than once will not cause errors,
3284 because the second time the file is included nothing will happen
3285 because the macro \c{MACROS_MAC} will already be defined.
3287 You can force a file to be included even if there is no \c{%include}
3288 directive that explicitly includes it, by using the \i\c{-p} option
3289 on the NASM command line (see \k{opt-p}).
3292 \S{pathsearch} \i\c{%pathsearch}: Search the Include Path
3294 The \c{%pathsearch} directive takes a single-line macro name and a
3295 filename, and declare or redefines the specified single-line macro to
3296 be the include-path-resolved version of the filename, if the file
3297 exists (otherwise, it is passed unchanged.)
3301 \c %pathsearch MyFoo "foo.bin"
3303 ... with \c{-Ibins/} in the include path may end up defining the macro
3304 \c{MyFoo} to be \c{"bins/foo.bin"}.
3307 \S{depend} \i\c{%depend}: Add Dependent Files
3309 The \c{%depend} directive takes a filename and adds it to the list of
3310 files to be emitted as dependency generation when the \c{-M} options
3311 and its relatives (see \k{opt-M}) are used. It produces no output.
3313 This is generally used in conjunction with \c{%pathsearch}. For
3314 example, a simplified version of the standard macro wrapper for the
3315 \c{INCBIN} directive looks like:
3317 \c %imacro incbin 1-2+ 0
3318 \c %pathsearch dep %1
3323 This first resolves the location of the file into the macro \c{dep},
3324 then adds it to the dependency lists, and finally issues the
3325 assembler-level \c{INCBIN} directive.
3328 \S{use} \i\c{%use}: Include Standard Macro Package
3330 The \c{%use} directive is similar to \c{%include}, but rather than
3331 including the contents of a file, it includes a named standard macro
3332 package. The standard macro packages are part of NASM, and are
3333 described in \k{macropkg}.
3335 Unlike the \c{%include} directive, package names for the \c{%use}
3336 directive do not require quotes, but quotes are permitted. In NASM
3337 2.04 and 2.05 the unquoted form would be macro-expanded; this is no
3338 longer true. Thus, the following lines are equivalent:
3343 Standard macro packages are protected from multiple inclusion. When a
3344 standard macro package is used, a testable single-line macro of the
3345 form \c{__USE_}\e{package}\c{__} is also defined, see \k{use_def}.
3347 \H{ctxstack} The \i{Context Stack}
3349 Having labels that are local to a macro definition is sometimes not
3350 quite powerful enough: sometimes you want to be able to share labels
3351 between several macro calls. An example might be a \c{REPEAT} ...
3352 \c{UNTIL} loop, in which the expansion of the \c{REPEAT} macro
3353 would need to be able to refer to a label which the \c{UNTIL} macro
3354 had defined. However, for such a macro you would also want to be
3355 able to nest these loops.
3357 NASM provides this level of power by means of a \e{context stack}.
3358 The preprocessor maintains a stack of \e{contexts}, each of which is
3359 characterized by a name. You add a new context to the stack using
3360 the \i\c{%push} directive, and remove one using \i\c{%pop}. You can
3361 define labels that are local to a particular context on the stack.
3364 \S{pushpop} \i\c{%push} and \i\c{%pop}: \I{creating
3365 contexts}\I{removing contexts}Creating and Removing Contexts
3367 The \c{%push} directive is used to create a new context and place it
3368 on the top of the context stack. \c{%push} takes an optional argument,
3369 which is the name of the context. For example:
3373 This pushes a new context called \c{foobar} on the stack. You can have
3374 several contexts on the stack with the same name: they can still be
3375 distinguished. If no name is given, the context is unnamed (this is
3376 normally used when both the \c{%push} and the \c{%pop} are inside a
3377 single macro definition.)
3379 The directive \c{%pop}, taking one optional argument, removes the top
3380 context from the context stack and destroys it, along with any
3381 labels associated with it. If an argument is given, it must match the
3382 name of the current context, otherwise it will issue an error.
3385 \S{ctxlocal} \i{Context-Local Labels}
3387 Just as the usage \c{%%foo} defines a label which is local to the
3388 particular macro call in which it is used, the usage \I{%$}\c{%$foo}
3389 is used to define a label which is local to the context on the top
3390 of the context stack. So the \c{REPEAT} and \c{UNTIL} example given
3391 above could be implemented by means of:
3407 and invoked by means of, for example,
3415 which would scan every fourth byte of a string in search of the byte
3418 If you need to define, or access, labels local to the context
3419 \e{below} the top one on the stack, you can use \I{%$$}\c{%$$foo}, or
3420 \c{%$$$foo} for the context below that, and so on.
3423 \S{ctxdefine} \i{Context-Local Single-Line Macros}
3425 NASM also allows you to define single-line macros which are local to
3426 a particular context, in just the same way:
3428 \c %define %$localmac 3
3430 will define the single-line macro \c{%$localmac} to be local to the
3431 top context on the stack. Of course, after a subsequent \c{%push},
3432 it can then still be accessed by the name \c{%$$localmac}.
3435 \S{ctxfallthrough} \i{Context Fall-Through Lookup}
3437 Context fall-through lookup (automatic searching of outer contexts)
3438 is a feature that was added in NASM version 0.98.03. Unfortunately,
3439 this feature is unintuitive and can result in buggy code that would
3440 have otherwise been prevented by NASM's error reporting. As a result,
3441 this feature has been \e{deprecated}. NASM version 2.09 will issue a
3442 warning when usage of this \e{deprecated} feature is detected. Starting
3443 with NASM version 2.10, usage of this \e{deprecated} feature will simply
3444 result in an \e{expression syntax error}.
3446 An example usage of this \e{deprecated} feature follows:
3450 \c %assign %$external 1
3452 \c %assign %$internal 1
3453 \c mov eax, %$external
3454 \c mov eax, %$internal
3459 As demonstrated, \c{%$external} is being defined in the \c{ctx1}
3460 context and referenced within the \c{ctx2} context. With context
3461 fall-through lookup, referencing an undefined context-local macro
3462 like this implicitly searches through all outer contexts until a match
3463 is made or isn't found in any context. As a result, \c{%$external}
3464 referenced within the \c{ctx2} context would implicitly use \c{%$external}
3465 as defined in \c{ctx1}. Most people would expect NASM to issue an error in
3466 this situation because \c{%$external} was never defined within \c{ctx2} and also
3467 isn't qualified with the proper context depth, \c{%$$external}.
3469 Here is a revision of the above example with proper context depth:
3473 \c %assign %$external 1
3475 \c %assign %$internal 1
3476 \c mov eax, %$$external
3477 \c mov eax, %$internal
3482 As demonstrated, \c{%$external} is still being defined in the \c{ctx1}
3483 context and referenced within the \c{ctx2} context. However, the
3484 reference to \c{%$external} within \c{ctx2} has been fully qualified with
3485 the proper context depth, \c{%$$external}, and thus is no longer ambiguous,
3486 unintuitive or erroneous.
3489 \S{ctxrepl} \i\c{%repl}: \I{renaming contexts}Renaming a Context
3491 If you need to change the name of the top context on the stack (in
3492 order, for example, to have it respond differently to \c{%ifctx}),
3493 you can execute a \c{%pop} followed by a \c{%push}; but this will
3494 have the side effect of destroying all context-local labels and
3495 macros associated with the context that was just popped.
3497 NASM provides the directive \c{%repl}, which \e{replaces} a context
3498 with a different name, without touching the associated macros and
3499 labels. So you could replace the destructive code
3504 with the non-destructive version \c{%repl newname}.
3507 \S{blockif} Example Use of the \i{Context Stack}: \i{Block IFs}
3509 This example makes use of almost all the context-stack features,
3510 including the conditional-assembly construct \i\c{%ifctx}, to
3511 implement a block IF statement as a set of macros.
3527 \c %error "expected `if' before `else'"
3541 \c %error "expected `if' or `else' before `endif'"
3546 This code is more robust than the \c{REPEAT} and \c{UNTIL} macros
3547 given in \k{ctxlocal}, because it uses conditional assembly to check
3548 that the macros are issued in the right order (for example, not
3549 calling \c{endif} before \c{if}) and issues a \c{%error} if they're
3552 In addition, the \c{endif} macro has to be able to cope with the two
3553 distinct cases of either directly following an \c{if}, or following
3554 an \c{else}. It achieves this, again, by using conditional assembly
3555 to do different things depending on whether the context on top of
3556 the stack is \c{if} or \c{else}.
3558 The \c{else} macro has to preserve the context on the stack, in
3559 order to have the \c{%$ifnot} referred to by the \c{if} macro be the
3560 same as the one defined by the \c{endif} macro, but has to change
3561 the context's name so that \c{endif} will know there was an
3562 intervening \c{else}. It does this by the use of \c{%repl}.
3564 A sample usage of these macros might look like:
3586 The block-\c{IF} macros handle nesting quite happily, by means of
3587 pushing another context, describing the inner \c{if}, on top of the
3588 one describing the outer \c{if}; thus \c{else} and \c{endif} always
3589 refer to the last unmatched \c{if} or \c{else}.
3592 \H{stackrel} \i{Stack Relative Preprocessor Directives}
3594 The following preprocessor directives provide a way to use
3595 labels to refer to local variables allocated on the stack.
3597 \b\c{%arg} (see \k{arg})
3599 \b\c{%stacksize} (see \k{stacksize})
3601 \b\c{%local} (see \k{local})
3604 \S{arg} \i\c{%arg} Directive
3606 The \c{%arg} directive is used to simplify the handling of
3607 parameters passed on the stack. Stack based parameter passing
3608 is used by many high level languages, including C, C++ and Pascal.
3610 While NASM has macros which attempt to duplicate this
3611 functionality (see \k{16cmacro}), the syntax is not particularly
3612 convenient to use and is not TASM compatible. Here is an example
3613 which shows the use of \c{%arg} without any external macros:
3617 \c %push mycontext ; save the current context
3618 \c %stacksize large ; tell NASM to use bp
3619 \c %arg i:word, j_ptr:word
3626 \c %pop ; restore original context
3628 This is similar to the procedure defined in \k{16cmacro} and adds
3629 the value in i to the value pointed to by j_ptr and returns the
3630 sum in the ax register. See \k{pushpop} for an explanation of
3631 \c{push} and \c{pop} and the use of context stacks.
3634 \S{stacksize} \i\c{%stacksize} Directive
3636 The \c{%stacksize} directive is used in conjunction with the
3637 \c{%arg} (see \k{arg}) and the \c{%local} (see \k{local}) directives.
3638 It tells NASM the default size to use for subsequent \c{%arg} and
3639 \c{%local} directives. The \c{%stacksize} directive takes one
3640 required argument which is one of \c{flat}, \c{flat64}, \c{large} or \c{small}.
3644 This form causes NASM to use stack-based parameter addressing
3645 relative to \c{ebp} and it assumes that a near form of call was used
3646 to get to this label (i.e. that \c{eip} is on the stack).
3648 \c %stacksize flat64
3650 This form causes NASM to use stack-based parameter addressing
3651 relative to \c{rbp} and it assumes that a near form of call was used
3652 to get to this label (i.e. that \c{rip} is on the stack).
3656 This form uses \c{bp} to do stack-based parameter addressing and
3657 assumes that a far form of call was used to get to this address
3658 (i.e. that \c{ip} and \c{cs} are on the stack).
3662 This form also uses \c{bp} to address stack parameters, but it is
3663 different from \c{large} because it also assumes that the old value
3664 of bp is pushed onto the stack (i.e. it expects an \c{ENTER}
3665 instruction). In other words, it expects that \c{bp}, \c{ip} and
3666 \c{cs} are on the top of the stack, underneath any local space which
3667 may have been allocated by \c{ENTER}. This form is probably most
3668 useful when used in combination with the \c{%local} directive
3672 \S{local} \i\c{%local} Directive
3674 The \c{%local} directive is used to simplify the use of local
3675 temporary stack variables allocated in a stack frame. Automatic
3676 local variables in C are an example of this kind of variable. The
3677 \c{%local} directive is most useful when used with the \c{%stacksize}
3678 (see \k{stacksize} and is also compatible with the \c{%arg} directive
3679 (see \k{arg}). It allows simplified reference to variables on the
3680 stack which have been allocated typically by using the \c{ENTER}
3682 \# (see \k{insENTER} for a description of that instruction).
3683 An example of its use is the following:
3687 \c %push mycontext ; save the current context
3688 \c %stacksize small ; tell NASM to use bp
3689 \c %assign %$localsize 0 ; see text for explanation
3690 \c %local old_ax:word, old_dx:word
3692 \c enter %$localsize,0 ; see text for explanation
3693 \c mov [old_ax],ax ; swap ax & bx
3694 \c mov [old_dx],dx ; and swap dx & cx
3699 \c leave ; restore old bp
3702 \c %pop ; restore original context
3704 The \c{%$localsize} variable is used internally by the
3705 \c{%local} directive and \e{must} be defined within the
3706 current context before the \c{%local} directive may be used.
3707 Failure to do so will result in one expression syntax error for
3708 each \c{%local} variable declared. It then may be used in
3709 the construction of an appropriately sized ENTER instruction
3710 as shown in the example.
3713 \H{pperror} Reporting \i{User-Defined Errors}: \i\c{%error}, \i\c{%warning}, \i\c{%fatal}
3715 The preprocessor directive \c{%error} will cause NASM to report an
3716 error if it occurs in assembled code. So if other users are going to
3717 try to assemble your source files, you can ensure that they define the
3718 right macros by means of code like this:
3723 \c ; do some different setup
3725 \c %error "Neither F1 nor F2 was defined."
3728 Then any user who fails to understand the way your code is supposed
3729 to be assembled will be quickly warned of their mistake, rather than
3730 having to wait until the program crashes on being run and then not
3731 knowing what went wrong.
3733 Similarly, \c{%warning} issues a warning, but allows assembly to continue:
3738 \c ; do some different setup
3740 \c %warning "Neither F1 nor F2 was defined, assuming F1."
3744 \c{%error} and \c{%warning} are issued only on the final assembly
3745 pass. This makes them safe to use in conjunction with tests that
3746 depend on symbol values.
3748 \c{%fatal} terminates assembly immediately, regardless of pass. This
3749 is useful when there is no point in continuing the assembly further,
3750 and doing so is likely just going to cause a spew of confusing error
3753 It is optional for the message string after \c{%error}, \c{%warning}
3754 or \c{%fatal} to be quoted. If it is \e{not}, then single-line macros
3755 are expanded in it, which can be used to display more information to
3756 the user. For example:
3759 \c %assign foo_over foo-64
3760 \c %error foo is foo_over bytes too large
3764 \H{otherpreproc} \i{Other Preprocessor Directives}
3766 NASM also has preprocessor directives which allow access to
3767 information from external sources. Currently they include:
3769 \b\c{%line} enables NASM to correctly handle the output of another
3770 preprocessor (see \k{line}).
3772 \b\c{%!} enables NASM to read in the value of an environment variable,
3773 which can then be used in your program (see \k{getenv}).
3775 \S{line} \i\c{%line} Directive
3777 The \c{%line} directive is used to notify NASM that the input line
3778 corresponds to a specific line number in another file. Typically
3779 this other file would be an original source file, with the current
3780 NASM input being the output of a pre-processor. The \c{%line}
3781 directive allows NASM to output messages which indicate the line
3782 number of the original source file, instead of the file that is being
3785 This preprocessor directive is not generally of use to programmers,
3786 by may be of interest to preprocessor authors. The usage of the
3787 \c{%line} preprocessor directive is as follows:
3789 \c %line nnn[+mmm] [filename]
3791 In this directive, \c{nnn} identifies the line of the original source
3792 file which this line corresponds to. \c{mmm} is an optional parameter
3793 which specifies a line increment value; each line of the input file
3794 read in is considered to correspond to \c{mmm} lines of the original
3795 source file. Finally, \c{filename} is an optional parameter which
3796 specifies the file name of the original source file.
3798 After reading a \c{%line} preprocessor directive, NASM will report
3799 all file name and line numbers relative to the values specified
3803 \S{getenv} \i\c{%!}\c{<env>}: Read an environment variable.
3805 The \c{%!<env>} directive makes it possible to read the value of an
3806 environment variable at assembly time. This could, for example, be used
3807 to store the contents of an environment variable into a string, which
3808 could be used at some other point in your code.
3810 For example, suppose that you have an environment variable \c{FOO}, and
3811 you want the contents of \c{FOO} to be embedded in your program. You
3812 could do that as follows:
3814 \c %defstr FOO %!FOO
3816 See \k{defstr} for notes on the \c{%defstr} directive.
3818 If the name of the environment variable contains non-identifier
3819 characters, you can use string quotes to surround the name of the
3820 variable, for example:
3822 \c %defstr C_colon %!'C:'
3825 \S{final} \i\c{%final} Directive
3827 The \c{%final} directive is used to delay preprocessing of a line
3828 until all other "normal" preprocessing is complete. Multiple
3829 \c{%final} directives are processed in the opposite order of their
3830 declaration, last one first and first one last.
3833 \H{comment} Comment Blocks: \i\c{%comment}
3835 The \c{%comment} and \c{%endcomment} directives are used to specify
3836 a block of commented (i.e. unprocessed) code/text. Everything between
3837 \c{%comment} and \c{%endcomment} will be ignored by the preprocessor.
3840 \c ; some code, text or data to be ignored
3844 \H{stdmac} \i{Standard Macros}
3846 NASM defines a set of standard macros, which are already defined
3847 when it starts to process any source file. If you really need a
3848 program to be assembled with no pre-defined macros, you can use the
3849 \i\c{%clear} directive to empty the preprocessor of everything but
3850 context-local preprocessor variables and single-line macros.
3852 Most \i{user-level assembler directives} (see \k{directive}) are
3853 implemented as macros which invoke primitive directives; these are
3854 described in \k{directive}. The rest of the standard macro set is
3858 \S{stdmacver} \i{NASM Version} Macros
3860 The single-line macros \i\c{__NASM_MAJOR__}, \i\c{__NASM_MINOR__},
3861 \i\c{__NASM_SUBMINOR__} and \i\c{___NASM_PATCHLEVEL__} expand to the
3862 major, minor, subminor and patch level parts of the \i{version
3863 number of NASM} being used. So, under NASM 0.98.32p1 for
3864 example, \c{__NASM_MAJOR__} would be defined to be 0, \c{__NASM_MINOR__}
3865 would be defined as 98, \c{__NASM_SUBMINOR__} would be defined to 32,
3866 and \c{___NASM_PATCHLEVEL__} would be defined as 1.
3868 Additionally, the macro \i\c{__NASM_SNAPSHOT__} is defined for
3869 automatically generated snapshot releases \e{only}.
3872 \S{stdmacverid} \i\c{__NASM_VERSION_ID__}: \i{NASM Version ID}
3874 The single-line macro \c{__NASM_VERSION_ID__} expands to a dword integer
3875 representing the full version number of the version of nasm being used.
3876 The value is the equivalent to \c{__NASM_MAJOR__}, \c{__NASM_MINOR__},
3877 \c{__NASM_SUBMINOR__} and \c{___NASM_PATCHLEVEL__} concatenated to
3878 produce a single doubleword. Hence, for 0.98.32p1, the returned number
3879 would be equivalent to:
3887 Note that the above lines are generate exactly the same code, the second
3888 line is used just to give an indication of the order that the separate
3889 values will be present in memory.
3892 \S{stdmacverstr} \i\c{__NASM_VER__}: \i{NASM Version string}
3894 The single-line macro \c{__NASM_VER__} expands to a string which defines
3895 the version number of nasm being used. So, under NASM 0.98.32 for example,
3904 \S{fileline} \i\c{__FILE__} and \i\c{__LINE__}: File Name and Line Number
3906 Like the C preprocessor, NASM allows the user to find out the file
3907 name and line number containing the current instruction. The macro
3908 \c{__FILE__} expands to a string constant giving the name of the
3909 current input file (which may change through the course of assembly
3910 if \c{%include} directives are used), and \c{__LINE__} expands to a
3911 numeric constant giving the current line number in the input file.
3913 These macros could be used, for example, to communicate debugging
3914 information to a macro, since invoking \c{__LINE__} inside a macro
3915 definition (either single-line or multi-line) will return the line
3916 number of the macro \e{call}, rather than \e{definition}. So to
3917 determine where in a piece of code a crash is occurring, for
3918 example, one could write a routine \c{stillhere}, which is passed a
3919 line number in \c{EAX} and outputs something like `line 155: still
3920 here'. You could then write a macro
3922 \c %macro notdeadyet 0
3931 and then pepper your code with calls to \c{notdeadyet} until you
3932 find the crash point.
3935 \S{bitsm} \i\c{__BITS__}: Current BITS Mode
3937 The \c{__BITS__} standard macro is updated every time that the BITS mode is
3938 set using the \c{BITS XX} or \c{[BITS XX]} directive, where XX is a valid mode
3939 number of 16, 32 or 64. \c{__BITS__} receives the specified mode number and
3940 makes it globally available. This can be very useful for those who utilize
3941 mode-dependent macros.
3943 \S{ofmtm} \i\c{__OUTPUT_FORMAT__}: Current Output Format
3945 The \c{__OUTPUT_FORMAT__} standard macro holds the current Output Format,
3946 as given by the \c{-f} option or NASM's default. Type \c{nasm -hf} for a
3949 \c %ifidn __OUTPUT_FORMAT__, win32
3950 \c %define NEWLINE 13, 10
3951 \c %elifidn __OUTPUT_FORMAT__, elf32
3952 \c %define NEWLINE 10
3956 \S{datetime} Assembly Date and Time Macros
3958 NASM provides a variety of macros that represent the timestamp of the
3961 \b The \i\c{__DATE__} and \i\c{__TIME__} macros give the assembly date and
3962 time as strings, in ISO 8601 format (\c{"YYYY-MM-DD"} and \c{"HH:MM:SS"},
3965 \b The \i\c{__DATE_NUM__} and \i\c{__TIME_NUM__} macros give the assembly
3966 date and time in numeric form; in the format \c{YYYYMMDD} and
3967 \c{HHMMSS} respectively.
3969 \b The \i\c{__UTC_DATE__} and \i\c{__UTC_TIME__} macros give the assembly
3970 date and time in universal time (UTC) as strings, in ISO 8601 format
3971 (\c{"YYYY-MM-DD"} and \c{"HH:MM:SS"}, respectively.) If the host
3972 platform doesn't provide UTC time, these macros are undefined.
3974 \b The \i\c{__UTC_DATE_NUM__} and \i\c{__UTC_TIME_NUM__} macros give the
3975 assembly date and time universal time (UTC) in numeric form; in the
3976 format \c{YYYYMMDD} and \c{HHMMSS} respectively. If the
3977 host platform doesn't provide UTC time, these macros are
3980 \b The \c{__POSIX_TIME__} macro is defined as a number containing the
3981 number of seconds since the POSIX epoch, 1 January 1970 00:00:00 UTC;
3982 excluding any leap seconds. This is computed using UTC time if
3983 available on the host platform, otherwise it is computed using the
3984 local time as if it was UTC.
3986 All instances of time and date macros in the same assembly session
3987 produce consistent output. For example, in an assembly session
3988 started at 42 seconds after midnight on January 1, 2010 in Moscow
3989 (timezone UTC+3) these macros would have the following values,
3990 assuming, of course, a properly configured environment with a correct
3993 \c __DATE__ "2010-01-01"
3994 \c __TIME__ "00:00:42"
3995 \c __DATE_NUM__ 20100101
3996 \c __TIME_NUM__ 000042
3997 \c __UTC_DATE__ "2009-12-31"
3998 \c __UTC_TIME__ "21:00:42"
3999 \c __UTC_DATE_NUM__ 20091231
4000 \c __UTC_TIME_NUM__ 210042
4001 \c __POSIX_TIME__ 1262293242
4004 \S{use_def} \I\c{__USE_*__}\c{__USE_}\e{package}\c{__}: Package
4007 When a standard macro package (see \k{macropkg}) is included with the
4008 \c{%use} directive (see \k{use}), a single-line macro of the form
4009 \c{__USE_}\e{package}\c{__} is automatically defined. This allows
4010 testing if a particular package is invoked or not.
4012 For example, if the \c{altreg} package is included (see
4013 \k{pkg_altreg}), then the macro \c{__USE_ALTREG__} is defined.
4016 \S{pass_macro} \i\c{__PASS__}: Assembly Pass
4018 The macro \c{__PASS__} is defined to be \c{1} on preparatory passes,
4019 and \c{2} on the final pass. In preprocess-only mode, it is set to
4020 \c{3}, and when running only to generate dependencies (due to the
4021 \c{-M} or \c{-MG} option, see \k{opt-M}) it is set to \c{0}.
4023 \e{Avoid using this macro if at all possible. It is tremendously easy
4024 to generate very strange errors by misusing it, and the semantics may
4025 change in future versions of NASM.}
4028 \S{struc} \i\c{STRUC} and \i\c{ENDSTRUC}: \i{Declaring Structure} Data Types
4030 The core of NASM contains no intrinsic means of defining data
4031 structures; instead, the preprocessor is sufficiently powerful that
4032 data structures can be implemented as a set of macros. The macros
4033 \c{STRUC} and \c{ENDSTRUC} are used to define a structure data type.
4035 \c{STRUC} takes one or two parameters. The first parameter is the name
4036 of the data type. The second, optional parameter is the base offset of
4037 the structure. The name of the data type is defined as a symbol with
4038 the value of the base offset, and the name of the data type with the
4039 suffix \c{_size} appended to it is defined as an \c{EQU} giving the
4040 size of the structure. Once \c{STRUC} has been issued, you are
4041 defining the structure, and should define fields using the \c{RESB}
4042 family of pseudo-instructions, and then invoke \c{ENDSTRUC} to finish
4045 For example, to define a structure called \c{mytype} containing a
4046 longword, a word, a byte and a string of bytes, you might code
4057 The above code defines six symbols: \c{mt_long} as 0 (the offset
4058 from the beginning of a \c{mytype} structure to the longword field),
4059 \c{mt_word} as 4, \c{mt_byte} as 6, \c{mt_str} as 7, \c{mytype_size}
4060 as 39, and \c{mytype} itself as zero.
4062 The reason why the structure type name is defined at zero by default
4063 is a side effect of allowing structures to work with the local label
4064 mechanism: if your structure members tend to have the same names in
4065 more than one structure, you can define the above structure like this:
4076 This defines the offsets to the structure fields as \c{mytype.long},
4077 \c{mytype.word}, \c{mytype.byte} and \c{mytype.str}.
4079 NASM, since it has no \e{intrinsic} structure support, does not
4080 support any form of period notation to refer to the elements of a
4081 structure once you have one (except the above local-label notation),
4082 so code such as \c{mov ax,[mystruc.mt_word]} is not valid.
4083 \c{mt_word} is a constant just like any other constant, so the
4084 correct syntax is \c{mov ax,[mystruc+mt_word]} or \c{mov
4085 ax,[mystruc+mytype.word]}.
4087 Sometimes you only have the address of the structure displaced by an
4088 offset. For example, consider this standard stack frame setup:
4094 In this case, you could access an element by subtracting the offset:
4096 \c mov [ebp - 40 + mytype.word], ax
4098 However, if you do not want to repeat this offset, you can use -40 as
4101 \c struc mytype, -40
4103 And access an element this way:
4105 \c mov [ebp + mytype.word], ax
4108 \S{istruc} \i\c{ISTRUC}, \i\c{AT} and \i\c{IEND}: Declaring
4109 \i{Instances of Structures}
4111 Having defined a structure type, the next thing you typically want
4112 to do is to declare instances of that structure in your data
4113 segment. NASM provides an easy way to do this in the \c{ISTRUC}
4114 mechanism. To declare a structure of type \c{mytype} in a program,
4115 you code something like this:
4120 \c at mt_long, dd 123456
4121 \c at mt_word, dw 1024
4122 \c at mt_byte, db 'x'
4123 \c at mt_str, db 'hello, world', 13, 10, 0
4127 The function of the \c{AT} macro is to make use of the \c{TIMES}
4128 prefix to advance the assembly position to the correct point for the
4129 specified structure field, and then to declare the specified data.
4130 Therefore the structure fields must be declared in the same order as
4131 they were specified in the structure definition.
4133 If the data to go in a structure field requires more than one source
4134 line to specify, the remaining source lines can easily come after
4135 the \c{AT} line. For example:
4137 \c at mt_str, db 123,134,145,156,167,178,189
4140 Depending on personal taste, you can also omit the code part of the
4141 \c{AT} line completely, and start the structure field on the next
4145 \c db 'hello, world'
4149 \S{align} \i\c{ALIGN} and \i\c{ALIGNB}: Data Alignment
4151 The \c{ALIGN} and \c{ALIGNB} macros provides a convenient way to
4152 align code or data on a word, longword, paragraph or other boundary.
4153 (Some assemblers call this directive \i\c{EVEN}.) The syntax of the
4154 \c{ALIGN} and \c{ALIGNB} macros is
4156 \c align 4 ; align on 4-byte boundary
4157 \c align 16 ; align on 16-byte boundary
4158 \c align 8,db 0 ; pad with 0s rather than NOPs
4159 \c align 4,resb 1 ; align to 4 in the BSS
4160 \c alignb 4 ; equivalent to previous line
4162 Both macros require their first argument to be a power of two; they
4163 both compute the number of additional bytes required to bring the
4164 length of the current section up to a multiple of that power of two,
4165 and then apply the \c{TIMES} prefix to their second argument to
4166 perform the alignment.
4168 If the second argument is not specified, the default for \c{ALIGN}
4169 is \c{NOP}, and the default for \c{ALIGNB} is \c{RESB 1}. So if the
4170 second argument is specified, the two macros are equivalent.
4171 Normally, you can just use \c{ALIGN} in code and data sections and
4172 \c{ALIGNB} in BSS sections, and never need the second argument
4173 except for special purposes.
4175 \c{ALIGN} and \c{ALIGNB}, being simple macros, perform no error
4176 checking: they cannot warn you if their first argument fails to be a
4177 power of two, or if their second argument generates more than one
4178 byte of code. In each of these cases they will silently do the wrong
4181 \c{ALIGNB} (or \c{ALIGN} with a second argument of \c{RESB 1}) can
4182 be used within structure definitions:
4199 This will ensure that the structure members are sensibly aligned
4200 relative to the base of the structure.
4202 A final caveat: \c{ALIGN} and \c{ALIGNB} work relative to the
4203 beginning of the \e{section}, not the beginning of the address space
4204 in the final executable. Aligning to a 16-byte boundary when the
4205 section you're in is only guaranteed to be aligned to a 4-byte
4206 boundary, for example, is a waste of effort. Again, NASM does not
4207 check that the section's alignment characteristics are sensible for
4208 the use of \c{ALIGN} or \c{ALIGNB}.
4210 Both \c{ALIGN} and \c{ALIGNB} do call \c{SECTALIGN} macro implicitly.
4211 See \k{sectalign} for details.
4213 See also the \c{smartalign} standard macro package, \k{pkg_smartalign}.
4216 \S{sectalign} \i\c{SECTALIGN}: Section Alignment
4218 The \c{SECTALIGN} macros provides a way to modify alignment attribute
4219 of output file section. Unlike the \c{align=} attribute (which is allowed
4220 at section definition only) the \c{SECTALIGN} macro may be used at any time.
4222 For example the directive
4226 sets the section alignment requirements to 16 bytes. Once increased it can
4227 not be decreased, the magnitude may grow only.
4229 Note that \c{ALIGN} (see \k{align}) calls the \c{SECTALIGN} macro implicitly
4230 so the active section alignment requirements may be updated. This is by default
4231 behaviour, if for some reason you want the \c{ALIGN} do not call \c{SECTALIGN}
4232 at all use the directive
4236 It is still possible to turn in on again by
4241 \C{macropkg} \i{Standard Macro Packages}
4243 The \i\c{%use} directive (see \k{use}) includes one of the standard
4244 macro packages included with the NASM distribution and compiled into
4245 the NASM binary. It operates like the \c{%include} directive (see
4246 \k{include}), but the included contents is provided by NASM itself.
4248 The names of standard macro packages are case insensitive, and can be
4252 \H{pkg_altreg} \i\c{altreg}: \i{Alternate Register Names}
4254 The \c{altreg} standard macro package provides alternate register
4255 names. It provides numeric register names for all registers (not just
4256 \c{R8}-\c{R15}), the Intel-defined aliases \c{R8L}-\c{R15L} for the
4257 low bytes of register (as opposed to the NASM/AMD standard names
4258 \c{R8B}-\c{R15B}), and the names \c{R0H}-\c{R3H} (by analogy with
4259 \c{R0L}-\c{R3L}) for \c{AH}, \c{CH}, \c{DH}, and \c{BH}.
4266 \c mov r0l,r3h ; mov al,bh
4272 \H{pkg_smartalign} \i\c{smartalign}\I{align, smart}: Smart \c{ALIGN} Macro
4274 The \c{smartalign} standard macro package provides for an \i\c{ALIGN}
4275 macro which is more powerful than the default (and
4276 backwards-compatible) one (see \k{align}). When the \c{smartalign}
4277 package is enabled, when \c{ALIGN} is used without a second argument,
4278 NASM will generate a sequence of instructions more efficient than a
4279 series of \c{NOP}. Furthermore, if the padding exceeds a specific
4280 threshold, then NASM will generate a jump over the entire padding
4283 The specific instructions generated can be controlled with the
4284 new \i\c{ALIGNMODE} macro. This macro takes two parameters: one mode,
4285 and an optional jump threshold override. If (for any reason) you need
4286 to turn off the jump completely just set jump threshold value to -1
4287 (or set it to \c{nojmp}). The following modes are possible:
4289 \b \c{generic}: Works on all x86 CPUs and should have reasonable
4290 performance. The default jump threshold is 8. This is the
4293 \b \c{nop}: Pad out with \c{NOP} instructions. The only difference
4294 compared to the standard \c{ALIGN} macro is that NASM can still jump
4295 over a large padding area. The default jump threshold is 16.
4297 \b \c{k7}: Optimize for the AMD K7 (Athlon/Althon XP). These
4298 instructions should still work on all x86 CPUs. The default jump
4301 \b \c{k8}: Optimize for the AMD K8 (Opteron/Althon 64). These
4302 instructions should still work on all x86 CPUs. The default jump
4305 \b \c{p6}: Optimize for Intel CPUs. This uses the long \c{NOP}
4306 instructions first introduced in Pentium Pro. This is incompatible
4307 with all CPUs of family 5 or lower, as well as some VIA CPUs and
4308 several virtualization solutions. The default jump threshold is 16.
4310 The macro \i\c{__ALIGNMODE__} is defined to contain the current
4311 alignment mode. A number of other macros beginning with \c{__ALIGN_}
4312 are used internally by this macro package.
4315 \H{pkg_fp} \i\c\{fp}: Floating-point macros
4317 This packages contains the following floating-point convenience macros:
4319 \c %define Inf __Infinity__
4320 \c %define NaN __QNaN__
4321 \c %define QNaN __QNaN__
4322 \c %define SNaN __SNaN__
4324 \c %define float8(x) __float8__(x)
4325 \c %define float16(x) __float16__(x)
4326 \c %define float32(x) __float32__(x)
4327 \c %define float64(x) __float64__(x)
4328 \c %define float80m(x) __float80m__(x)
4329 \c %define float80e(x) __float80e__(x)
4330 \c %define float128l(x) __float128l__(x)
4331 \c %define float128h(x) __float128h__(x)
4334 \C{directive} \i{Assembler Directives}
4336 NASM, though it attempts to avoid the bureaucracy of assemblers like
4337 MASM and TASM, is nevertheless forced to support a \e{few}
4338 directives. These are described in this chapter.
4340 NASM's directives come in two types: \I{user-level
4341 directives}\e{user-level} directives and \I{primitive
4342 directives}\e{primitive} directives. Typically, each directive has a
4343 user-level form and a primitive form. In almost all cases, we
4344 recommend that users use the user-level forms of the directives,
4345 which are implemented as macros which call the primitive forms.
4347 Primitive directives are enclosed in square brackets; user-level
4350 In addition to the universal directives described in this chapter,
4351 each object file format can optionally supply extra directives in
4352 order to control particular features of that file format. These
4353 \I{format-specific directives}\e{format-specific} directives are
4354 documented along with the formats that implement them, in \k{outfmt}.
4357 \H{bits} \i\c{BITS}: Specifying Target \i{Processor Mode}
4359 The \c{BITS} directive specifies whether NASM should generate code
4360 \I{16-bit mode, versus 32-bit mode}designed to run on a processor
4361 operating in 16-bit mode, 32-bit mode or 64-bit mode. The syntax is
4362 \c{BITS XX}, where XX is 16, 32 or 64.
4364 In most cases, you should not need to use \c{BITS} explicitly. The
4365 \c{aout}, \c{coff}, \c{elf}, \c{macho}, \c{win32} and \c{win64}
4366 object formats, which are designed for use in 32-bit or 64-bit
4367 operating systems, all cause NASM to select 32-bit or 64-bit mode,
4368 respectively, by default. The \c{obj} object format allows you
4369 to specify each segment you define as either \c{USE16} or \c{USE32},
4370 and NASM will set its operating mode accordingly, so the use of the
4371 \c{BITS} directive is once again unnecessary.
4373 The most likely reason for using the \c{BITS} directive is to write
4374 32-bit or 64-bit code in a flat binary file; this is because the \c{bin}
4375 output format defaults to 16-bit mode in anticipation of it being
4376 used most frequently to write DOS \c{.COM} programs, DOS \c{.SYS}
4377 device drivers and boot loader software.
4379 You do \e{not} need to specify \c{BITS 32} merely in order to use
4380 32-bit instructions in a 16-bit DOS program; if you do, the
4381 assembler will generate incorrect code because it will be writing
4382 code targeted at a 32-bit platform, to be run on a 16-bit one.
4384 When NASM is in \c{BITS 16} mode, instructions which use 32-bit
4385 data are prefixed with an 0x66 byte, and those referring to 32-bit
4386 addresses have an 0x67 prefix. In \c{BITS 32} mode, the reverse is
4387 true: 32-bit instructions require no prefixes, whereas instructions
4388 using 16-bit data need an 0x66 and those working on 16-bit addresses
4391 When NASM is in \c{BITS 64} mode, most instructions operate the same
4392 as they do for \c{BITS 32} mode. However, there are 8 more general and
4393 SSE registers, and 16-bit addressing is no longer supported.
4395 The default address size is 64 bits; 32-bit addressing can be selected
4396 with the 0x67 prefix. The default operand size is still 32 bits,
4397 however, and the 0x66 prefix selects 16-bit operand size. The \c{REX}
4398 prefix is used both to select 64-bit operand size, and to access the
4399 new registers. NASM automatically inserts REX prefixes when
4402 When the \c{REX} prefix is used, the processor does not know how to
4403 address the AH, BH, CH or DH (high 8-bit legacy) registers. Instead,
4404 it is possible to access the the low 8-bits of the SP, BP SI and DI
4405 registers as SPL, BPL, SIL and DIL, respectively; but only when the
4408 The \c{BITS} directive has an exactly equivalent primitive form,
4409 \c{[BITS 16]}, \c{[BITS 32]} and \c{[BITS 64]}. The user-level form is
4410 a macro which has no function other than to call the primitive form.
4412 Note that the space is neccessary, e.g. \c{BITS32} will \e{not} work!
4414 \S{USE16 & USE32} \i\c{USE16} & \i\c{USE32}: Aliases for BITS
4416 The `\c{USE16}' and `\c{USE32}' directives can be used in place of
4417 `\c{BITS 16}' and `\c{BITS 32}', for compatibility with other assemblers.
4420 \H{default} \i\c{DEFAULT}: Change the assembler defaults
4422 The \c{DEFAULT} directive changes the assembler defaults. Normally,
4423 NASM defaults to a mode where the programmer is expected to explicitly
4424 specify most features directly. However, this is occationally
4425 obnoxious, as the explicit form is pretty much the only one one wishes
4428 Currently, the only \c{DEFAULT} that is settable is whether or not
4429 registerless instructions in 64-bit mode are \c{RIP}-relative or not.
4430 By default, they are absolute unless overridden with the \i\c{REL}
4431 specifier (see \k{effaddr}). However, if \c{DEFAULT REL} is
4432 specified, \c{REL} is default, unless overridden with the \c{ABS}
4433 specifier, \e{except when used with an FS or GS segment override}.
4435 The special handling of \c{FS} and \c{GS} overrides are due to the
4436 fact that these registers are generally used as thread pointers or
4437 other special functions in 64-bit mode, and generating
4438 \c{RIP}-relative addresses would be extremely confusing.
4440 \c{DEFAULT REL} is disabled with \c{DEFAULT ABS}.
4442 \H{section} \i\c{SECTION} or \i\c{SEGMENT}: Changing and \i{Defining
4445 \I{changing sections}\I{switching between sections}The \c{SECTION}
4446 directive (\c{SEGMENT} is an exactly equivalent synonym) changes
4447 which section of the output file the code you write will be
4448 assembled into. In some object file formats, the number and names of
4449 sections are fixed; in others, the user may make up as many as they
4450 wish. Hence \c{SECTION} may sometimes give an error message, or may
4451 define a new section, if you try to switch to a section that does
4454 The Unix object formats, and the \c{bin} object format (but see
4455 \k{multisec}, all support
4456 the \i{standardized section names} \c{.text}, \c{.data} and \c{.bss}
4457 for the code, data and uninitialized-data sections. The \c{obj}
4458 format, by contrast, does not recognize these section names as being
4459 special, and indeed will strip off the leading period of any section
4463 \S{sectmac} The \i\c{__SECT__} Macro
4465 The \c{SECTION} directive is unusual in that its user-level form
4466 functions differently from its primitive form. The primitive form,
4467 \c{[SECTION xyz]}, simply switches the current target section to the
4468 one given. The user-level form, \c{SECTION xyz}, however, first
4469 defines the single-line macro \c{__SECT__} to be the primitive
4470 \c{[SECTION]} directive which it is about to issue, and then issues
4471 it. So the user-level directive
4475 expands to the two lines
4477 \c %define __SECT__ [SECTION .text]
4480 Users may find it useful to make use of this in their own macros.
4481 For example, the \c{writefile} macro defined in \k{mlmacgre} can be
4482 usefully rewritten in the following more sophisticated form:
4484 \c %macro writefile 2+
4494 \c mov cx,%%endstr-%%str
4501 This form of the macro, once passed a string to output, first
4502 switches temporarily to the data section of the file, using the
4503 primitive form of the \c{SECTION} directive so as not to modify
4504 \c{__SECT__}. It then declares its string in the data section, and
4505 then invokes \c{__SECT__} to switch back to \e{whichever} section
4506 the user was previously working in. It thus avoids the need, in the
4507 previous version of the macro, to include a \c{JMP} instruction to
4508 jump over the data, and also does not fail if, in a complicated
4509 \c{OBJ} format module, the user could potentially be assembling the
4510 code in any of several separate code sections.
4513 \H{absolute} \i\c{ABSOLUTE}: Defining Absolute Labels
4515 The \c{ABSOLUTE} directive can be thought of as an alternative form
4516 of \c{SECTION}: it causes the subsequent code to be directed at no
4517 physical section, but at the hypothetical section starting at the
4518 given absolute address. The only instructions you can use in this
4519 mode are the \c{RESB} family.
4521 \c{ABSOLUTE} is used as follows:
4529 This example describes a section of the PC BIOS data area, at
4530 segment address 0x40: the above code defines \c{kbuf_chr} to be
4531 0x1A, \c{kbuf_free} to be 0x1C, and \c{kbuf} to be 0x1E.
4533 The user-level form of \c{ABSOLUTE}, like that of \c{SECTION},
4534 redefines the \i\c{__SECT__} macro when it is invoked.
4536 \i\c{STRUC} and \i\c{ENDSTRUC} are defined as macros which use
4537 \c{ABSOLUTE} (and also \c{__SECT__}).
4539 \c{ABSOLUTE} doesn't have to take an absolute constant as an
4540 argument: it can take an expression (actually, a \i{critical
4541 expression}: see \k{crit}) and it can be a value in a segment. For
4542 example, a TSR can re-use its setup code as run-time BSS like this:
4544 \c org 100h ; it's a .COM program
4546 \c jmp setup ; setup code comes last
4548 \c ; the resident part of the TSR goes here
4550 \c ; now write the code that installs the TSR here
4554 \c runtimevar1 resw 1
4555 \c runtimevar2 resd 20
4559 This defines some variables `on top of' the setup code, so that
4560 after the setup has finished running, the space it took up can be
4561 re-used as data storage for the running TSR. The symbol `tsr_end'
4562 can be used to calculate the total size of the part of the TSR that
4563 needs to be made resident.
4566 \H{extern} \i\c{EXTERN}: \i{Importing Symbols} from Other Modules
4568 \c{EXTERN} is similar to the MASM directive \c{EXTRN} and the C
4569 keyword \c{extern}: it is used to declare a symbol which is not
4570 defined anywhere in the module being assembled, but is assumed to be
4571 defined in some other module and needs to be referred to by this
4572 one. Not every object-file format can support external variables:
4573 the \c{bin} format cannot.
4575 The \c{EXTERN} directive takes as many arguments as you like. Each
4576 argument is the name of a symbol:
4579 \c extern _sscanf,_fscanf
4581 Some object-file formats provide extra features to the \c{EXTERN}
4582 directive. In all cases, the extra features are used by suffixing a
4583 colon to the symbol name followed by object-format specific text.
4584 For example, the \c{obj} format allows you to declare that the
4585 default segment base of an external should be the group \c{dgroup}
4586 by means of the directive
4588 \c extern _variable:wrt dgroup
4590 The primitive form of \c{EXTERN} differs from the user-level form
4591 only in that it can take only one argument at a time: the support
4592 for multiple arguments is implemented at the preprocessor level.
4594 You can declare the same variable as \c{EXTERN} more than once: NASM
4595 will quietly ignore the second and later redeclarations. You can't
4596 declare a variable as \c{EXTERN} as well as something else, though.
4599 \H{global} \i\c{GLOBAL}: \i{Exporting Symbols} to Other Modules
4601 \c{GLOBAL} is the other end of \c{EXTERN}: if one module declares a
4602 symbol as \c{EXTERN} and refers to it, then in order to prevent
4603 linker errors, some other module must actually \e{define} the
4604 symbol and declare it as \c{GLOBAL}. Some assemblers use the name
4605 \i\c{PUBLIC} for this purpose.
4607 The \c{GLOBAL} directive applying to a symbol must appear \e{before}
4608 the definition of the symbol.
4610 \c{GLOBAL} uses the same syntax as \c{EXTERN}, except that it must
4611 refer to symbols which \e{are} defined in the same module as the
4612 \c{GLOBAL} directive. For example:
4618 \c{GLOBAL}, like \c{EXTERN}, allows object formats to define private
4619 extensions by means of a colon. The \c{elf} object format, for
4620 example, lets you specify whether global data items are functions or
4623 \c global hashlookup:function, hashtable:data
4625 Like \c{EXTERN}, the primitive form of \c{GLOBAL} differs from the
4626 user-level form only in that it can take only one argument at a
4630 \H{common} \i\c{COMMON}: Defining Common Data Areas
4632 The \c{COMMON} directive is used to declare \i\e{common variables}.
4633 A common variable is much like a global variable declared in the
4634 uninitialized data section, so that
4638 is similar in function to
4645 The difference is that if more than one module defines the same
4646 common variable, then at link time those variables will be
4647 \e{merged}, and references to \c{intvar} in all modules will point
4648 at the same piece of memory.
4650 Like \c{GLOBAL} and \c{EXTERN}, \c{COMMON} supports object-format
4651 specific extensions. For example, the \c{obj} format allows common
4652 variables to be NEAR or FAR, and the \c{elf} format allows you to
4653 specify the alignment requirements of a common variable:
4655 \c common commvar 4:near ; works in OBJ
4656 \c common intarray 100:4 ; works in ELF: 4 byte aligned
4658 Once again, like \c{EXTERN} and \c{GLOBAL}, the primitive form of
4659 \c{COMMON} differs from the user-level form only in that it can take
4660 only one argument at a time.
4663 \H{CPU} \i\c{CPU}: Defining CPU Dependencies
4665 The \i\c{CPU} directive restricts assembly to those instructions which
4666 are available on the specified CPU.
4670 \b\c{CPU 8086} Assemble only 8086 instruction set
4672 \b\c{CPU 186} Assemble instructions up to the 80186 instruction set
4674 \b\c{CPU 286} Assemble instructions up to the 286 instruction set
4676 \b\c{CPU 386} Assemble instructions up to the 386 instruction set
4678 \b\c{CPU 486} 486 instruction set
4680 \b\c{CPU 586} Pentium instruction set
4682 \b\c{CPU PENTIUM} Same as 586
4684 \b\c{CPU 686} P6 instruction set
4686 \b\c{CPU PPRO} Same as 686
4688 \b\c{CPU P2} Same as 686
4690 \b\c{CPU P3} Pentium III (Katmai) instruction sets
4692 \b\c{CPU KATMAI} Same as P3
4694 \b\c{CPU P4} Pentium 4 (Willamette) instruction set
4696 \b\c{CPU WILLAMETTE} Same as P4
4698 \b\c{CPU PRESCOTT} Prescott instruction set
4700 \b\c{CPU X64} x86-64 (x64/AMD64/Intel 64) instruction set
4702 \b\c{CPU IA64} IA64 CPU (in x86 mode) instruction set
4704 All options are case insensitive. All instructions will be selected
4705 only if they apply to the selected CPU or lower. By default, all
4706 instructions are available.
4709 \H{FLOAT} \i\c{FLOAT}: Handling of \I{floating-point, constants}floating-point constants
4711 By default, floating-point constants are rounded to nearest, and IEEE
4712 denormals are supported. The following options can be set to alter
4715 \b\c{FLOAT DAZ} Flush denormals to zero
4717 \b\c{FLOAT NODAZ} Do not flush denormals to zero (default)
4719 \b\c{FLOAT NEAR} Round to nearest (default)
4721 \b\c{FLOAT UP} Round up (toward +Infinity)
4723 \b\c{FLOAT DOWN} Round down (toward -Infinity)
4725 \b\c{FLOAT ZERO} Round toward zero
4727 \b\c{FLOAT DEFAULT} Restore default settings
4729 The standard macros \i\c{__FLOAT_DAZ__}, \i\c{__FLOAT_ROUND__}, and
4730 \i\c{__FLOAT__} contain the current state, as long as the programmer
4731 has avoided the use of the brackeded primitive form, (\c{[FLOAT]}).
4733 \c{__FLOAT__} contains the full set of floating-point settings; this
4734 value can be saved away and invoked later to restore the setting.
4737 \C{outfmt} \i{Output Formats}
4739 NASM is a portable assembler, designed to be able to compile on any
4740 ANSI C-supporting platform and produce output to run on a variety of
4741 Intel x86 operating systems. For this reason, it has a large number
4742 of available output formats, selected using the \i\c{-f} option on
4743 the NASM \i{command line}. Each of these formats, along with its
4744 extensions to the base NASM syntax, is detailed in this chapter.
4746 As stated in \k{opt-o}, NASM chooses a \i{default name} for your
4747 output file based on the input file name and the chosen output
4748 format. This will be generated by removing the \i{extension}
4749 (\c{.asm}, \c{.s}, or whatever you like to use) from the input file
4750 name, and substituting an extension defined by the output format.
4751 The extensions are given with each format below.
4754 \H{binfmt} \i\c{bin}: \i{Flat-Form Binary}\I{pure binary} Output
4756 The \c{bin} format does not produce object files: it generates
4757 nothing in the output file except the code you wrote. Such `pure
4758 binary' files are used by \i{MS-DOS}: \i\c{.COM} executables and
4759 \i\c{.SYS} device drivers are pure binary files. Pure binary output
4760 is also useful for \i{operating system} and \i{boot loader}
4763 The \c{bin} format supports \i{multiple section names}. For details of
4764 how NASM handles sections in the \c{bin} format, see \k{multisec}.
4766 Using the \c{bin} format puts NASM by default into 16-bit mode (see
4767 \k{bits}). In order to use \c{bin} to write 32-bit or 64-bit code,
4768 such as an OS kernel, you need to explicitly issue the \I\c{BITS}\c{BITS 32}
4769 or \I\c{BITS}\c{BITS 64} directive.
4771 \c{bin} has no default output file name extension: instead, it
4772 leaves your file name as it is once the original extension has been
4773 removed. Thus, the default is for NASM to assemble \c{binprog.asm}
4774 into a binary file called \c{binprog}.
4777 \S{org} \i\c{ORG}: Binary File \i{Program Origin}
4779 The \c{bin} format provides an additional directive to the list
4780 given in \k{directive}: \c{ORG}. The function of the \c{ORG}
4781 directive is to specify the origin address which NASM will assume
4782 the program begins at when it is loaded into memory.
4784 For example, the following code will generate the longword
4791 Unlike the \c{ORG} directive provided by MASM-compatible assemblers,
4792 which allows you to jump around in the object file and overwrite
4793 code you have already generated, NASM's \c{ORG} does exactly what
4794 the directive says: \e{origin}. Its sole function is to specify one
4795 offset which is added to all internal address references within the
4796 section; it does not permit any of the trickery that MASM's version
4797 does. See \k{proborg} for further comments.
4800 \S{binseg} \c{bin} Extensions to the \c{SECTION}
4801 Directive\I{SECTION, bin extensions to}
4803 The \c{bin} output format extends the \c{SECTION} (or \c{SEGMENT})
4804 directive to allow you to specify the alignment requirements of
4805 segments. This is done by appending the \i\c{ALIGN} qualifier to the
4806 end of the section-definition line. For example,
4808 \c section .data align=16
4810 switches to the section \c{.data} and also specifies that it must be
4811 aligned on a 16-byte boundary.
4813 The parameter to \c{ALIGN} specifies how many low bits of the
4814 section start address must be forced to zero. The alignment value
4815 given may be any power of two.\I{section alignment, in
4816 bin}\I{segment alignment, in bin}\I{alignment, in bin sections}
4819 \S{multisec} \i{Multisection}\I{bin, multisection} Support for the \c{bin} Format
4821 The \c{bin} format allows the use of multiple sections, of arbitrary names,
4822 besides the "known" \c{.text}, \c{.data}, and \c{.bss} names.
4824 \b Sections may be designated \i\c{progbits} or \i\c{nobits}. Default
4825 is \c{progbits} (except \c{.bss}, which defaults to \c{nobits},
4828 \b Sections can be aligned at a specified boundary following the previous
4829 section with \c{align=}, or at an arbitrary byte-granular position with
4832 \b Sections can be given a virtual start address, which will be used
4833 for the calculation of all memory references within that section
4836 \b Sections can be ordered using \i\c{follows=}\c{<section>} or
4837 \i\c{vfollows=}\c{<section>} as an alternative to specifying an explicit
4840 \b Arguments to \c{org}, \c{start}, \c{vstart}, and \c{align=} are
4841 critical expressions. See \k{crit}. E.g. \c{align=(1 << ALIGN_SHIFT)}
4842 - \c{ALIGN_SHIFT} must be defined before it is used here.
4844 \b Any code which comes before an explicit \c{SECTION} directive
4845 is directed by default into the \c{.text} section.
4847 \b If an \c{ORG} statement is not given, \c{ORG 0} is used
4850 \b The \c{.bss} section will be placed after the last \c{progbits}
4851 section, unless \c{start=}, \c{vstart=}, \c{follows=}, or \c{vfollows=}
4854 \b All sections are aligned on dword boundaries, unless a different
4855 alignment has been specified.
4857 \b Sections may not overlap.
4859 \b NASM creates the \c{section.<secname>.start} for each section,
4860 which may be used in your code.
4862 \S{map}\i{Map Files}
4864 Map files can be generated in \c{-f bin} format by means of the \c{[map]}
4865 option. Map types of \c{all} (default), \c{brief}, \c{sections}, \c{segments},
4866 or \c{symbols} may be specified. Output may be directed to \c{stdout}
4867 (default), \c{stderr}, or a specified file. E.g.
4868 \c{[map symbols myfile.map]}. No "user form" exists, the square
4869 brackets must be used.
4872 \H{ithfmt} \i\c{ith}: \i{Intel Hex} Output
4874 The \c{ith} file format produces Intel hex-format files. Just as the
4875 \c{bin} format, this is a flat memory image format with no support for
4876 relocation or linking. It is usually used with ROM programmers and
4879 All extensions supported by the \c{bin} file format is also supported by
4880 the \c{ith} file format.
4882 \c{ith} provides a default output file-name extension of \c{.ith}.
4885 \H{srecfmt} \i\c{srec}: \i{Motorola S-Records} Output
4887 The \c{srec} file format produces Motorola S-records files. Just as the
4888 \c{bin} format, this is a flat memory image format with no support for
4889 relocation or linking. It is usually used with ROM programmers and
4892 All extensions supported by the \c{bin} file format is also supported by
4893 the \c{srec} file format.
4895 \c{srec} provides a default output file-name extension of \c{.srec}.
4898 \H{objfmt} \i\c{obj}: \i{Microsoft OMF}\I{OMF} Object Files
4900 The \c{obj} file format (NASM calls it \c{obj} rather than \c{omf}
4901 for historical reasons) is the one produced by \i{MASM} and
4902 \i{TASM}, which is typically fed to 16-bit DOS linkers to produce
4903 \i\c{.EXE} files. It is also the format used by \i{OS/2}.
4905 \c{obj} provides a default output file-name extension of \c{.obj}.
4907 \c{obj} is not exclusively a 16-bit format, though: NASM has full
4908 support for the 32-bit extensions to the format. In particular,
4909 32-bit \c{obj} format files are used by \i{Borland's Win32
4910 compilers}, instead of using Microsoft's newer \i\c{win32} object
4913 The \c{obj} format does not define any special segment names: you
4914 can call your segments anything you like. Typical names for segments
4915 in \c{obj} format files are \c{CODE}, \c{DATA} and \c{BSS}.
4917 If your source file contains code before specifying an explicit
4918 \c{SEGMENT} directive, then NASM will invent its own segment called
4919 \i\c{__NASMDEFSEG} for you.
4921 When you define a segment in an \c{obj} file, NASM defines the
4922 segment name as a symbol as well, so that you can access the segment
4923 address of the segment. So, for example:
4932 \c mov ax,data ; get segment address of data
4933 \c mov ds,ax ; and move it into DS
4934 \c inc word [dvar] ; now this reference will work
4937 The \c{obj} format also enables the use of the \i\c{SEG} and
4938 \i\c{WRT} operators, so that you can write code which does things
4943 \c mov ax,seg foo ; get preferred segment of foo
4945 \c mov ax,data ; a different segment
4947 \c mov ax,[ds:foo] ; this accesses `foo'
4948 \c mov [es:foo wrt data],bx ; so does this
4951 \S{objseg} \c{obj} Extensions to the \c{SEGMENT}
4952 Directive\I{SEGMENT, obj extensions to}
4954 The \c{obj} output format extends the \c{SEGMENT} (or \c{SECTION})
4955 directive to allow you to specify various properties of the segment
4956 you are defining. This is done by appending extra qualifiers to the
4957 end of the segment-definition line. For example,
4959 \c segment code private align=16
4961 defines the segment \c{code}, but also declares it to be a private
4962 segment, and requires that the portion of it described in this code
4963 module must be aligned on a 16-byte boundary.
4965 The available qualifiers are:
4967 \b \i\c{PRIVATE}, \i\c{PUBLIC}, \i\c{COMMON} and \i\c{STACK} specify
4968 the combination characteristics of the segment. \c{PRIVATE} segments
4969 do not get combined with any others by the linker; \c{PUBLIC} and
4970 \c{STACK} segments get concatenated together at link time; and
4971 \c{COMMON} segments all get overlaid on top of each other rather
4972 than stuck end-to-end.
4974 \b \i\c{ALIGN} is used, as shown above, to specify how many low bits
4975 of the segment start address must be forced to zero. The alignment
4976 value given may be any power of two from 1 to 4096; in reality, the
4977 only values supported are 1, 2, 4, 16, 256 and 4096, so if 8 is
4978 specified it will be rounded up to 16, and 32, 64 and 128 will all
4979 be rounded up to 256, and so on. Note that alignment to 4096-byte
4980 boundaries is a \i{PharLap} extension to the format and may not be
4981 supported by all linkers.\I{section alignment, in OBJ}\I{segment
4982 alignment, in OBJ}\I{alignment, in OBJ sections}
4984 \b \i\c{CLASS} can be used to specify the segment class; this feature
4985 indicates to the linker that segments of the same class should be
4986 placed near each other in the output file. The class name can be any
4987 word, e.g. \c{CLASS=CODE}.
4989 \b \i\c{OVERLAY}, like \c{CLASS}, is specified with an arbitrary word
4990 as an argument, and provides overlay information to an
4991 overlay-capable linker.
4993 \b Segments can be declared as \i\c{USE16} or \i\c{USE32}, which has
4994 the effect of recording the choice in the object file and also
4995 ensuring that NASM's default assembly mode when assembling in that
4996 segment is 16-bit or 32-bit respectively.
4998 \b When writing \i{OS/2} object files, you should declare 32-bit
4999 segments as \i\c{FLAT}, which causes the default segment base for
5000 anything in the segment to be the special group \c{FLAT}, and also
5001 defines the group if it is not already defined.
5003 \b The \c{obj} file format also allows segments to be declared as
5004 having a pre-defined absolute segment address, although no linkers
5005 are currently known to make sensible use of this feature;
5006 nevertheless, NASM allows you to declare a segment such as
5007 \c{SEGMENT SCREEN ABSOLUTE=0xB800} if you need to. The \i\c{ABSOLUTE}
5008 and \c{ALIGN} keywords are mutually exclusive.
5010 NASM's default segment attributes are \c{PUBLIC}, \c{ALIGN=1}, no
5011 class, no overlay, and \c{USE16}.
5014 \S{group} \i\c{GROUP}: Defining Groups of Segments\I{segments, groups of}
5016 The \c{obj} format also allows segments to be grouped, so that a
5017 single segment register can be used to refer to all the segments in
5018 a group. NASM therefore supplies the \c{GROUP} directive, whereby
5027 \c ; some uninitialized data
5029 \c group dgroup data bss
5031 which will define a group called \c{dgroup} to contain the segments
5032 \c{data} and \c{bss}. Like \c{SEGMENT}, \c{GROUP} causes the group
5033 name to be defined as a symbol, so that you can refer to a variable
5034 \c{var} in the \c{data} segment as \c{var wrt data} or as \c{var wrt
5035 dgroup}, depending on which segment value is currently in your
5038 If you just refer to \c{var}, however, and \c{var} is declared in a
5039 segment which is part of a group, then NASM will default to giving
5040 you the offset of \c{var} from the beginning of the \e{group}, not
5041 the \e{segment}. Therefore \c{SEG var}, also, will return the group
5042 base rather than the segment base.
5044 NASM will allow a segment to be part of more than one group, but
5045 will generate a warning if you do this. Variables declared in a
5046 segment which is part of more than one group will default to being
5047 relative to the first group that was defined to contain the segment.
5049 A group does not have to contain any segments; you can still make
5050 \c{WRT} references to a group which does not contain the variable
5051 you are referring to. OS/2, for example, defines the special group
5052 \c{FLAT} with no segments in it.
5055 \S{uppercase} \i\c{UPPERCASE}: Disabling Case Sensitivity in Output
5057 Although NASM itself is \i{case sensitive}, some OMF linkers are
5058 not; therefore it can be useful for NASM to output single-case
5059 object files. The \c{UPPERCASE} format-specific directive causes all
5060 segment, group and symbol names that are written to the object file
5061 to be forced to upper case just before being written. Within a
5062 source file, NASM is still case-sensitive; but the object file can
5063 be written entirely in upper case if desired.
5065 \c{UPPERCASE} is used alone on a line; it requires no parameters.
5068 \S{import} \i\c{IMPORT}: Importing DLL Symbols\I{DLL symbols,
5069 importing}\I{symbols, importing from DLLs}
5071 The \c{IMPORT} format-specific directive defines a symbol to be
5072 imported from a DLL, for use if you are writing a DLL's \i{import
5073 library} in NASM. You still need to declare the symbol as \c{EXTERN}
5074 as well as using the \c{IMPORT} directive.
5076 The \c{IMPORT} directive takes two required parameters, separated by
5077 white space, which are (respectively) the name of the symbol you
5078 wish to import and the name of the library you wish to import it
5081 \c import WSAStartup wsock32.dll
5083 A third optional parameter gives the name by which the symbol is
5084 known in the library you are importing it from, in case this is not
5085 the same as the name you wish the symbol to be known by to your code
5086 once you have imported it. For example:
5088 \c import asyncsel wsock32.dll WSAAsyncSelect
5091 \S{export} \i\c{EXPORT}: Exporting DLL Symbols\I{DLL symbols,
5092 exporting}\I{symbols, exporting from DLLs}
5094 The \c{EXPORT} format-specific directive defines a global symbol to
5095 be exported as a DLL symbol, for use if you are writing a DLL in
5096 NASM. You still need to declare the symbol as \c{GLOBAL} as well as
5097 using the \c{EXPORT} directive.
5099 \c{EXPORT} takes one required parameter, which is the name of the
5100 symbol you wish to export, as it was defined in your source file. An
5101 optional second parameter (separated by white space from the first)
5102 gives the \e{external} name of the symbol: the name by which you
5103 wish the symbol to be known to programs using the DLL. If this name
5104 is the same as the internal name, you may leave the second parameter
5107 Further parameters can be given to define attributes of the exported
5108 symbol. These parameters, like the second, are separated by white
5109 space. If further parameters are given, the external name must also
5110 be specified, even if it is the same as the internal name. The
5111 available attributes are:
5113 \b \c{resident} indicates that the exported name is to be kept
5114 resident by the system loader. This is an optimisation for
5115 frequently used symbols imported by name.
5117 \b \c{nodata} indicates that the exported symbol is a function which
5118 does not make use of any initialized data.
5120 \b \c{parm=NNN}, where \c{NNN} is an integer, sets the number of
5121 parameter words for the case in which the symbol is a call gate
5122 between 32-bit and 16-bit segments.
5124 \b An attribute which is just a number indicates that the symbol
5125 should be exported with an identifying number (ordinal), and gives
5131 \c export myfunc TheRealMoreFormalLookingFunctionName
5132 \c export myfunc myfunc 1234 ; export by ordinal
5133 \c export myfunc myfunc resident parm=23 nodata
5136 \S{dotdotstart} \i\c{..start}: Defining the \i{Program Entry
5139 \c{OMF} linkers require exactly one of the object files being linked to
5140 define the program entry point, where execution will begin when the
5141 program is run. If the object file that defines the entry point is
5142 assembled using NASM, you specify the entry point by declaring the
5143 special symbol \c{..start} at the point where you wish execution to
5147 \S{objextern} \c{obj} Extensions to the \c{EXTERN}
5148 Directive\I{EXTERN, obj extensions to}
5150 If you declare an external symbol with the directive
5154 then references such as \c{mov ax,foo} will give you the offset of
5155 \c{foo} from its preferred segment base (as specified in whichever
5156 module \c{foo} is actually defined in). So to access the contents of
5157 \c{foo} you will usually need to do something like
5159 \c mov ax,seg foo ; get preferred segment base
5160 \c mov es,ax ; move it into ES
5161 \c mov ax,[es:foo] ; and use offset `foo' from it
5163 This is a little unwieldy, particularly if you know that an external
5164 is going to be accessible from a given segment or group, say
5165 \c{dgroup}. So if \c{DS} already contained \c{dgroup}, you could
5168 \c mov ax,[foo wrt dgroup]
5170 However, having to type this every time you want to access \c{foo}
5171 can be a pain; so NASM allows you to declare \c{foo} in the
5174 \c extern foo:wrt dgroup
5176 This form causes NASM to pretend that the preferred segment base of
5177 \c{foo} is in fact \c{dgroup}; so the expression \c{seg foo} will
5178 now return \c{dgroup}, and the expression \c{foo} is equivalent to
5181 This \I{default-WRT mechanism}default-\c{WRT} mechanism can be used
5182 to make externals appear to be relative to any group or segment in
5183 your program. It can also be applied to common variables: see
5187 \S{objcommon} \c{obj} Extensions to the \c{COMMON}
5188 Directive\I{COMMON, obj extensions to}
5190 The \c{obj} format allows common variables to be either near\I{near
5191 common variables} or far\I{far common variables}; NASM allows you to
5192 specify which your variables should be by the use of the syntax
5194 \c common nearvar 2:near ; `nearvar' is a near common
5195 \c common farvar 10:far ; and `farvar' is far
5197 Far common variables may be greater in size than 64Kb, and so the
5198 OMF specification says that they are declared as a number of
5199 \e{elements} of a given size. So a 10-byte far common variable could
5200 be declared as ten one-byte elements, five two-byte elements, two
5201 five-byte elements or one ten-byte element.
5203 Some \c{OMF} linkers require the \I{element size, in common
5204 variables}\I{common variables, element size}element size, as well as
5205 the variable size, to match when resolving common variables declared
5206 in more than one module. Therefore NASM must allow you to specify
5207 the element size on your far common variables. This is done by the
5210 \c common c_5by2 10:far 5 ; two five-byte elements
5211 \c common c_2by5 10:far 2 ; five two-byte elements
5213 If no element size is specified, the default is 1. Also, the \c{FAR}
5214 keyword is not required when an element size is specified, since
5215 only far commons may have element sizes at all. So the above
5216 declarations could equivalently be
5218 \c common c_5by2 10:5 ; two five-byte elements
5219 \c common c_2by5 10:2 ; five two-byte elements
5221 In addition to these extensions, the \c{COMMON} directive in \c{obj}
5222 also supports default-\c{WRT} specification like \c{EXTERN} does
5223 (explained in \k{objextern}). So you can also declare things like
5225 \c common foo 10:wrt dgroup
5226 \c common bar 16:far 2:wrt data
5227 \c common baz 24:wrt data:6
5230 \H{win32fmt} \i\c{win32}: Microsoft Win32 Object Files
5232 The \c{win32} output format generates Microsoft Win32 object files,
5233 suitable for passing to Microsoft linkers such as \i{Visual C++}.
5234 Note that Borland Win32 compilers do not use this format, but use
5235 \c{obj} instead (see \k{objfmt}).
5237 \c{win32} provides a default output file-name extension of \c{.obj}.
5239 Note that although Microsoft say that Win32 object files follow the
5240 \c{COFF} (Common Object File Format) standard, the object files produced
5241 by Microsoft Win32 compilers are not compatible with COFF linkers
5242 such as DJGPP's, and vice versa. This is due to a difference of
5243 opinion over the precise semantics of PC-relative relocations. To
5244 produce COFF files suitable for DJGPP, use NASM's \c{coff} output
5245 format; conversely, the \c{coff} format does not produce object
5246 files that Win32 linkers can generate correct output from.
5249 \S{win32sect} \c{win32} Extensions to the \c{SECTION}
5250 Directive\I{SECTION, win32 extensions to}
5252 Like the \c{obj} format, \c{win32} allows you to specify additional
5253 information on the \c{SECTION} directive line, to control the type
5254 and properties of sections you declare. Section types and properties
5255 are generated automatically by NASM for the \i{standard section names}
5256 \c{.text}, \c{.data} and \c{.bss}, but may still be overridden by
5259 The available qualifiers are:
5261 \b \c{code}, or equivalently \c{text}, defines the section to be a
5262 code section. This marks the section as readable and executable, but
5263 not writable, and also indicates to the linker that the type of the
5266 \b \c{data} and \c{bss} define the section to be a data section,
5267 analogously to \c{code}. Data sections are marked as readable and
5268 writable, but not executable. \c{data} declares an initialized data
5269 section, whereas \c{bss} declares an uninitialized data section.
5271 \b \c{rdata} declares an initialized data section that is readable
5272 but not writable. Microsoft compilers use this section to place
5275 \b \c{info} defines the section to be an \i{informational section},
5276 which is not included in the executable file by the linker, but may
5277 (for example) pass information \e{to} the linker. For example,
5278 declaring an \c{info}-type section called \i\c{.drectve} causes the
5279 linker to interpret the contents of the section as command-line
5282 \b \c{align=}, used with a trailing number as in \c{obj}, gives the
5283 \I{section alignment, in win32}\I{alignment, in win32
5284 sections}alignment requirements of the section. The maximum you may
5285 specify is 64: the Win32 object file format contains no means to
5286 request a greater section alignment than this. If alignment is not
5287 explicitly specified, the defaults are 16-byte alignment for code
5288 sections, 8-byte alignment for rdata sections and 4-byte alignment
5289 for data (and BSS) sections.
5290 Informational sections get a default alignment of 1 byte (no
5291 alignment), though the value does not matter.
5293 The defaults assumed by NASM if you do not specify the above
5296 \c section .text code align=16
5297 \c section .data data align=4
5298 \c section .rdata rdata align=8
5299 \c section .bss bss align=4
5301 Any other section name is treated by default like \c{.text}.
5303 \S{win32safeseh} \c{win32}: Safe Structured Exception Handling
5305 Among other improvements in Windows XP SP2 and Windows Server 2003
5306 Microsoft has introduced concept of "safe structured exception
5307 handling." General idea is to collect handlers' entry points in
5308 designated read-only table and have alleged entry point verified
5309 against this table prior exception control is passed to the handler. In
5310 order for an executable module to be equipped with such "safe exception
5311 handler table," all object modules on linker command line has to comply
5312 with certain criteria. If one single module among them does not, then
5313 the table in question is omitted and above mentioned run-time checks
5314 will not be performed for application in question. Table omission is by
5315 default silent and therefore can be easily overlooked. One can instruct
5316 linker to refuse to produce binary without such table by passing
5317 \c{/safeseh} command line option.
5319 Without regard to this run-time check merits it's natural to expect
5320 NASM to be capable of generating modules suitable for \c{/safeseh}
5321 linking. From developer's viewpoint the problem is two-fold:
5323 \b how to adapt modules not deploying exception handlers of their own;
5325 \b how to adapt/develop modules utilizing custom exception handling;
5327 Former can be easily achieved with any NASM version by adding following
5328 line to source code:
5332 As of version 2.03 NASM adds this absolute symbol automatically. If
5333 it's not already present to be precise. I.e. if for whatever reason
5334 developer would choose to assign another value in source file, it would
5335 still be perfectly possible.
5337 Registering custom exception handler on the other hand requires certain
5338 "magic." As of version 2.03 additional directive is implemented,
5339 \c{safeseh}, which instructs the assembler to produce appropriately
5340 formatted input data for above mentioned "safe exception handler
5341 table." Its typical use would be:
5344 \c extern _MessageBoxA@16
5345 \c %if __NASM_VERSION_ID__ >= 0x02030000
5346 \c safeseh handler ; register handler as "safe handler"
5349 \c push DWORD 1 ; MB_OKCANCEL
5350 \c push DWORD caption
5353 \c call _MessageBoxA@16
5354 \c sub eax,1 ; incidentally suits as return value
5355 \c ; for exception handler
5359 \c push DWORD handler
5360 \c push DWORD [fs:0]
5361 \c mov DWORD [fs:0],esp ; engage exception handler
5363 \c mov eax,DWORD[eax] ; cause exception
5364 \c pop DWORD [fs:0] ; disengage exception handler
5367 \c text: db 'OK to rethrow, CANCEL to generate core dump',0
5368 \c caption:db 'SEGV',0
5370 \c section .drectve info
5371 \c db '/defaultlib:user32.lib /defaultlib:msvcrt.lib '
5373 As you might imagine, it's perfectly possible to produce .exe binary
5374 with "safe exception handler table" and yet engage unregistered
5375 exception handler. Indeed, handler is engaged by simply manipulating
5376 \c{[fs:0]} location at run-time, something linker has no power over,
5377 run-time that is. It should be explicitly mentioned that such failure
5378 to register handler's entry point with \c{safeseh} directive has
5379 undesired side effect at run-time. If exception is raised and
5380 unregistered handler is to be executed, the application is abruptly
5381 terminated without any notification whatsoever. One can argue that
5382 system could at least have logged some kind "non-safe exception
5383 handler in x.exe at address n" message in event log, but no, literally
5384 no notification is provided and user is left with no clue on what
5385 caused application failure.
5387 Finally, all mentions of linker in this paragraph refer to Microsoft
5388 linker version 7.x and later. Presence of \c{@feat.00} symbol and input
5389 data for "safe exception handler table" causes no backward
5390 incompatibilities and "safeseh" modules generated by NASM 2.03 and
5391 later can still be linked by earlier versions or non-Microsoft linkers.
5394 \H{win64fmt} \i\c{win64}: Microsoft Win64 Object Files
5396 The \c{win64} output format generates Microsoft Win64 object files,
5397 which is nearly 100% identical to the \c{win32} object format (\k{win32fmt})
5398 with the exception that it is meant to target 64-bit code and the x86-64
5399 platform altogether. This object file is used exactly the same as the \c{win32}
5400 object format (\k{win32fmt}), in NASM, with regard to this exception.
5402 \S{win64pic} \c{win64}: Writing Position-Independent Code
5404 While \c{REL} takes good care of RIP-relative addressing, there is one
5405 aspect that is easy to overlook for a Win64 programmer: indirect
5406 references. Consider a switch dispatch table:
5408 \c jmp QWORD[dsptch+rax*8]
5414 Even novice Win64 assembler programmer will soon realize that the code
5415 is not 64-bit savvy. Most notably linker will refuse to link it with
5416 "\c{'ADDR32' relocation to '.text' invalid without
5417 /LARGEADDRESSAWARE:NO}". So [s]he will have to split jmp instruction as
5420 \c lea rbx,[rel dsptch]
5421 \c jmp QWORD[rbx+rax*8]
5423 What happens behind the scene is that effective address in \c{lea} is
5424 encoded relative to instruction pointer, or in perfectly
5425 position-independent manner. But this is only part of the problem!
5426 Trouble is that in .dll context \c{caseN} relocations will make their
5427 way to the final module and might have to be adjusted at .dll load
5428 time. To be specific when it can't be loaded at preferred address. And
5429 when this occurs, pages with such relocations will be rendered private
5430 to current process, which kind of undermines the idea of sharing .dll.
5431 But no worry, it's trivial to fix:
5433 \c lea rbx,[rel dsptch]
5434 \c add rbx,QWORD[rbx+rax*8]
5437 \c dsptch: dq case0-dsptch
5441 NASM version 2.03 and later provides another alternative, \c{wrt
5442 ..imagebase} operator, which returns offset from base address of the
5443 current image, be it .exe or .dll module, therefore the name. For those
5444 acquainted with PE-COFF format base address denotes start of
5445 \c{IMAGE_DOS_HEADER} structure. Here is how to implement switch with
5446 these image-relative references:
5448 \c lea rbx,[rel dsptch]
5449 \c mov eax,DWORD[rbx+rax*4]
5450 \c sub rbx,dsptch wrt ..imagebase
5454 \c dsptch: dd case0 wrt ..imagebase
5455 \c dd case1 wrt ..imagebase
5457 One can argue that the operator is redundant. Indeed, snippet before
5458 last works just fine with any NASM version and is not even Windows
5459 specific... The real reason for implementing \c{wrt ..imagebase} will
5460 become apparent in next paragraph.
5462 It should be noted that \c{wrt ..imagebase} is defined as 32-bit
5465 \c dd label wrt ..imagebase ; ok
5466 \c dq label wrt ..imagebase ; bad
5467 \c mov eax,label wrt ..imagebase ; ok
5468 \c mov rax,label wrt ..imagebase ; bad
5470 \S{win64seh} \c{win64}: Structured Exception Handling
5472 Structured exception handing in Win64 is completely different matter
5473 from Win32. Upon exception program counter value is noted, and
5474 linker-generated table comprising start and end addresses of all the
5475 functions [in given executable module] is traversed and compared to the
5476 saved program counter. Thus so called \c{UNWIND_INFO} structure is
5477 identified. If it's not found, then offending subroutine is assumed to
5478 be "leaf" and just mentioned lookup procedure is attempted for its
5479 caller. In Win64 leaf function is such function that does not call any
5480 other function \e{nor} modifies any Win64 non-volatile registers,
5481 including stack pointer. The latter ensures that it's possible to
5482 identify leaf function's caller by simply pulling the value from the
5485 While majority of subroutines written in assembler are not calling any
5486 other function, requirement for non-volatile registers' immutability
5487 leaves developer with not more than 7 registers and no stack frame,
5488 which is not necessarily what [s]he counted with. Customarily one would
5489 meet the requirement by saving non-volatile registers on stack and
5490 restoring them upon return, so what can go wrong? If [and only if] an
5491 exception is raised at run-time and no \c{UNWIND_INFO} structure is
5492 associated with such "leaf" function, the stack unwind procedure will
5493 expect to find caller's return address on the top of stack immediately
5494 followed by its frame. Given that developer pushed caller's
5495 non-volatile registers on stack, would the value on top point at some
5496 code segment or even addressable space? Well, developer can attempt
5497 copying caller's return address to the top of stack and this would
5498 actually work in some very specific circumstances. But unless developer
5499 can guarantee that these circumstances are always met, it's more
5500 appropriate to assume worst case scenario, i.e. stack unwind procedure
5501 going berserk. Relevant question is what happens then? Application is
5502 abruptly terminated without any notification whatsoever. Just like in
5503 Win32 case, one can argue that system could at least have logged
5504 "unwind procedure went berserk in x.exe at address n" in event log, but
5505 no, no trace of failure is left.
5507 Now, when we understand significance of the \c{UNWIND_INFO} structure,
5508 let's discuss what's in it and/or how it's processed. First of all it
5509 is checked for presence of reference to custom language-specific
5510 exception handler. If there is one, then it's invoked. Depending on the
5511 return value, execution flow is resumed (exception is said to be
5512 "handled"), \e{or} rest of \c{UNWIND_INFO} structure is processed as
5513 following. Beside optional reference to custom handler, it carries
5514 information about current callee's stack frame and where non-volatile
5515 registers are saved. Information is detailed enough to be able to
5516 reconstruct contents of caller's non-volatile registers upon call to
5517 current callee. And so caller's context is reconstructed, and then
5518 unwind procedure is repeated, i.e. another \c{UNWIND_INFO} structure is
5519 associated, this time, with caller's instruction pointer, which is then
5520 checked for presence of reference to language-specific handler, etc.
5521 The procedure is recursively repeated till exception is handled. As
5522 last resort system "handles" it by generating memory core dump and
5523 terminating the application.
5525 As for the moment of this writing NASM unfortunately does not
5526 facilitate generation of above mentioned detailed information about
5527 stack frame layout. But as of version 2.03 it implements building
5528 blocks for generating structures involved in stack unwinding. As
5529 simplest example, here is how to deploy custom exception handler for
5534 \c extern MessageBoxA
5540 \c mov r9,1 ; MB_OKCANCEL
5542 \c sub eax,1 ; incidentally suits as return value
5543 \c ; for exception handler
5549 \c mov rax,QWORD[rax] ; cause exception
5552 \c text: db 'OK to rethrow, CANCEL to generate core dump',0
5553 \c caption:db 'SEGV',0
5555 \c section .pdata rdata align=4
5556 \c dd main wrt ..imagebase
5557 \c dd main_end wrt ..imagebase
5558 \c dd xmain wrt ..imagebase
5559 \c section .xdata rdata align=8
5560 \c xmain: db 9,0,0,0
5561 \c dd handler wrt ..imagebase
5562 \c section .drectve info
5563 \c db '/defaultlib:user32.lib /defaultlib:msvcrt.lib '
5565 What you see in \c{.pdata} section is element of the "table comprising
5566 start and end addresses of function" along with reference to associated
5567 \c{UNWIND_INFO} structure. And what you see in \c{.xdata} section is
5568 \c{UNWIND_INFO} structure describing function with no frame, but with
5569 designated exception handler. References are \e{required} to be
5570 image-relative (which is the real reason for implementing \c{wrt
5571 ..imagebase} operator). It should be noted that \c{rdata align=n}, as
5572 well as \c{wrt ..imagebase}, are optional in these two segments'
5573 contexts, i.e. can be omitted. Latter means that \e{all} 32-bit
5574 references, not only above listed required ones, placed into these two
5575 segments turn out image-relative. Why is it important to understand?
5576 Developer is allowed to append handler-specific data to \c{UNWIND_INFO}
5577 structure, and if [s]he adds a 32-bit reference, then [s]he will have
5578 to remember to adjust its value to obtain the real pointer.
5580 As already mentioned, in Win64 terms leaf function is one that does not
5581 call any other function \e{nor} modifies any non-volatile register,
5582 including stack pointer. But it's not uncommon that assembler
5583 programmer plans to utilize every single register and sometimes even
5584 have variable stack frame. Is there anything one can do with bare
5585 building blocks? I.e. besides manually composing fully-fledged
5586 \c{UNWIND_INFO} structure, which would surely be considered
5587 error-prone? Yes, there is. Recall that exception handler is called
5588 first, before stack layout is analyzed. As it turned out, it's
5589 perfectly possible to manipulate current callee's context in custom
5590 handler in manner that permits further stack unwinding. General idea is
5591 that handler would not actually "handle" the exception, but instead
5592 restore callee's context, as it was at its entry point and thus mimic
5593 leaf function. In other words, handler would simply undertake part of
5594 unwinding procedure. Consider following example:
5597 \c mov rax,rsp ; copy rsp to volatile register
5598 \c push r15 ; save non-volatile registers
5601 \c mov r11,rsp ; prepare variable stack frame
5604 \c mov QWORD[r11],rax ; check for exceptions
5605 \c mov rsp,r11 ; allocate stack frame
5606 \c mov QWORD[rsp],rax ; save original rsp value
5609 \c mov r11,QWORD[rsp] ; pull original rsp value
5610 \c mov rbp,QWORD[r11-24]
5611 \c mov rbx,QWORD[r11-16]
5612 \c mov r15,QWORD[r11-8]
5613 \c mov rsp,r11 ; destroy frame
5616 The keyword is that up to \c{magic_point} original \c{rsp} value
5617 remains in chosen volatile register and no non-volatile register,
5618 except for \c{rsp}, is modified. While past \c{magic_point} \c{rsp}
5619 remains constant till the very end of the \c{function}. In this case
5620 custom language-specific exception handler would look like this:
5622 \c EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
5623 \c CONTEXT *context,DISPATCHER_CONTEXT *disp)
5625 \c if (context->Rip<(ULONG64)magic_point)
5626 \c rsp = (ULONG64 *)context->Rax;
5628 \c { rsp = ((ULONG64 **)context->Rsp)[0];
5629 \c context->Rbp = rsp[-3];
5630 \c context->Rbx = rsp[-2];
5631 \c context->R15 = rsp[-1];
5633 \c context->Rsp = (ULONG64)rsp;
5635 \c memcpy (disp->ContextRecord,context,sizeof(CONTEXT));
5636 \c RtlVirtualUnwind(UNW_FLAG_NHANDLER,disp->ImageBase,
5637 \c dips->ControlPc,disp->FunctionEntry,disp->ContextRecord,
5638 \c &disp->HandlerData,&disp->EstablisherFrame,NULL);
5639 \c return ExceptionContinueSearch;
5642 As custom handler mimics leaf function, corresponding \c{UNWIND_INFO}
5643 structure does not have to contain any information about stack frame
5646 \H{cofffmt} \i\c{coff}: \i{Common Object File Format}
5648 The \c{coff} output type produces \c{COFF} object files suitable for
5649 linking with the \i{DJGPP} linker.
5651 \c{coff} provides a default output file-name extension of \c{.o}.
5653 The \c{coff} format supports the same extensions to the \c{SECTION}
5654 directive as \c{win32} does, except that the \c{align} qualifier and
5655 the \c{info} section type are not supported.
5657 \H{machofmt} \I{Mach-O}\i\c{macho32} and \i\c{macho64}: \i{Mach Object File Format}
5659 The \c{macho32} and \c{macho64} output formts produces \c{Mach-O}
5660 object files suitable for linking with the \i{MacOS X} linker.
5661 \i\c{macho} is a synonym for \c{macho32}.
5663 \c{macho} provides a default output file-name extension of \c{.o}.
5665 \H{elffmt} \i\c{elf32}, \i\c{elf64}, \i\c{elfx32}: \I{ELF}\I{linux, elf}\i{Executable and Linkable
5666 Format} Object Files
5668 The \c{elf32}, \c{elf64} and \c{elfx32} output formats generate
5669 \c{ELF32 and ELF64} (Executable and Linkable Format) object files, as
5670 used by Linux as well as \i{Unix System V}, including \i{Solaris x86},
5671 \i{UnixWare} and \i{SCO Unix}. \c{elf} provides a default output
5672 file-name extension of \c{.o}. \c{elf} is a synonym for \c{elf32}.
5674 The \c{elfx32} format is used for the \i{x32} ABI, which is a 32-bit
5675 ABI with the CPU in 64-bit mode.
5677 \S{abisect} ELF specific directive \i\c{osabi}
5679 The ELF header specifies the application binary interface for the target operating system (OSABI).
5680 This field can be set by using the \c{osabi} directive with the numeric value (0-255) of the target
5681 system. If this directive is not used, the default value will be "UNIX System V ABI" (0) which will work on
5682 most systems which support ELF.
5684 \S{elfsect} \c{elf} Extensions to the \c{SECTION}
5685 Directive\I{SECTION, elf extensions to}
5687 Like the \c{obj} format, \c{elf} allows you to specify additional
5688 information on the \c{SECTION} directive line, to control the type
5689 and properties of sections you declare. Section types and properties
5690 are generated automatically by NASM for the \i{standard section
5691 names}, but may still be
5692 overridden by these qualifiers.
5694 The available qualifiers are:
5696 \b \i\c{alloc} defines the section to be one which is loaded into
5697 memory when the program is run. \i\c{noalloc} defines it to be one
5698 which is not, such as an informational or comment section.
5700 \b \i\c{exec} defines the section to be one which should have execute
5701 permission when the program is run. \i\c{noexec} defines it as one
5704 \b \i\c{write} defines the section to be one which should be writable
5705 when the program is run. \i\c{nowrite} defines it as one which should
5708 \b \i\c{progbits} defines the section to be one with explicit contents
5709 stored in the object file: an ordinary code or data section, for
5710 example, \i\c{nobits} defines the section to be one with no explicit
5711 contents given, such as a BSS section.
5713 \b \c{align=}, used with a trailing number as in \c{obj}, gives the
5714 \I{section alignment, in elf}\I{alignment, in elf sections}alignment
5715 requirements of the section.
5717 \b \i\c{tls} defines the section to be one which contains
5718 thread local variables.
5720 The defaults assumed by NASM if you do not specify the above
5723 \I\c{.text} \I\c{.rodata} \I\c{.lrodata} \I\c{.data} \I\c{.ldata}
5724 \I\c{.bss} \I\c{.lbss} \I\c{.tdata} \I\c{.tbss} \I\c\{.comment}
5726 \c section .text progbits alloc exec nowrite align=16
5727 \c section .rodata progbits alloc noexec nowrite align=4
5728 \c section .lrodata progbits alloc noexec nowrite align=4
5729 \c section .data progbits alloc noexec write align=4
5730 \c section .ldata progbits alloc noexec write align=4
5731 \c section .bss nobits alloc noexec write align=4
5732 \c section .lbss nobits alloc noexec write align=4
5733 \c section .tdata progbits alloc noexec write align=4 tls
5734 \c section .tbss nobits alloc noexec write align=4 tls
5735 \c section .comment progbits noalloc noexec nowrite align=1
5736 \c section other progbits alloc noexec nowrite align=1
5738 (Any section name other than those in the above table
5739 is treated by default like \c{other} in the above table.
5740 Please note that section names are case sensitive.)
5743 \S{elfwrt} \i{Position-Independent Code}\I{PIC}: \c{elf} Special
5744 Symbols and \i\c{WRT}
5746 The \c{ELF} specification contains enough features to allow
5747 position-independent code (PIC) to be written, which makes \i{ELF
5748 shared libraries} very flexible. However, it also means NASM has to
5749 be able to generate a variety of ELF specific relocation types in ELF
5750 object files, if it is to be an assembler which can write PIC.
5752 Since \c{ELF} does not support segment-base references, the \c{WRT}
5753 operator is not used for its normal purpose; therefore NASM's
5754 \c{elf} output format makes use of \c{WRT} for a different purpose,
5755 namely the PIC-specific \I{relocations, PIC-specific}relocation
5758 \c{elf} defines five special symbols which you can use as the
5759 right-hand side of the \c{WRT} operator to obtain PIC relocation
5760 types. They are \i\c{..gotpc}, \i\c{..gotoff}, \i\c{..got},
5761 \i\c{..plt} and \i\c{..sym}. Their functions are summarized here:
5763 \b Referring to the symbol marking the global offset table base
5764 using \c{wrt ..gotpc} will end up giving the distance from the
5765 beginning of the current section to the global offset table.
5766 (\i\c{_GLOBAL_OFFSET_TABLE_} is the standard symbol name used to
5767 refer to the \i{GOT}.) So you would then need to add \i\c{$$} to the
5768 result to get the real address of the GOT.
5770 \b Referring to a location in one of your own sections using \c{wrt
5771 ..gotoff} will give the distance from the beginning of the GOT to
5772 the specified location, so that adding on the address of the GOT
5773 would give the real address of the location you wanted.
5775 \b Referring to an external or global symbol using \c{wrt ..got}
5776 causes the linker to build an entry \e{in} the GOT containing the
5777 address of the symbol, and the reference gives the distance from the
5778 beginning of the GOT to the entry; so you can add on the address of
5779 the GOT, load from the resulting address, and end up with the
5780 address of the symbol.
5782 \b Referring to a procedure name using \c{wrt ..plt} causes the
5783 linker to build a \i{procedure linkage table} entry for the symbol,
5784 and the reference gives the address of the \i{PLT} entry. You can
5785 only use this in contexts which would generate a PC-relative
5786 relocation normally (i.e. as the destination for \c{CALL} or
5787 \c{JMP}), since ELF contains no relocation type to refer to PLT
5790 \b Referring to a symbol name using \c{wrt ..sym} causes NASM to
5791 write an ordinary relocation, but instead of making the relocation
5792 relative to the start of the section and then adding on the offset
5793 to the symbol, it will write a relocation record aimed directly at
5794 the symbol in question. The distinction is a necessary one due to a
5795 peculiarity of the dynamic linker.
5797 A fuller explanation of how to use these relocation types to write
5798 shared libraries entirely in NASM is given in \k{picdll}.
5800 \S{elftls} \i{Thread Local Storage}\I{TLS}: \c{elf} Special
5801 Symbols and \i\c{WRT}
5803 \b In ELF32 mode, referring to an external or global symbol using
5804 \c{wrt ..tlsie} \I\c{..tlsie}
5805 causes the linker to build an entry \e{in} the GOT containing the
5806 offset of the symbol within the TLS block, so you can access the value
5807 of the symbol with code such as:
5809 \c mov eax,[tid wrt ..tlsie]
5813 \b In ELF64 or ELFx32 mode, referring to an external or global symbol using
5814 \c{wrt ..gottpoff} \I\c{..gottpoff}
5815 causes the linker to build an entry \e{in} the GOT containing the
5816 offset of the symbol within the TLS block, so you can access the value
5817 of the symbol with code such as:
5819 \c mov rax,[rel tid wrt ..gottpoff]
5823 \S{elfglob} \c{elf} Extensions to the \c{GLOBAL} Directive\I{GLOBAL,
5824 elf extensions to}\I{GLOBAL, aoutb extensions to}
5826 \c{ELF} object files can contain more information about a global symbol
5827 than just its address: they can contain the \I{symbol sizes,
5828 specifying}\I{size, of symbols}size of the symbol and its \I{symbol
5829 types, specifying}\I{type, of symbols}type as well. These are not
5830 merely debugger conveniences, but are actually necessary when the
5831 program being written is a \i{shared library}. NASM therefore
5832 supports some extensions to the \c{GLOBAL} directive, allowing you
5833 to specify these features.
5835 You can specify whether a global variable is a function or a data
5836 object by suffixing the name with a colon and the word
5837 \i\c{function} or \i\c{data}. (\i\c{object} is a synonym for
5838 \c{data}.) For example:
5840 \c global hashlookup:function, hashtable:data
5842 exports the global symbol \c{hashlookup} as a function and
5843 \c{hashtable} as a data object.
5845 Optionally, you can control the ELF visibility of the symbol. Just
5846 add one of the visibility keywords: \i\c{default}, \i\c{internal},
5847 \i\c{hidden}, or \i\c{protected}. The default is \i\c{default} of
5848 course. For example, to make \c{hashlookup} hidden:
5850 \c global hashlookup:function hidden
5852 You can also specify the size of the data associated with the
5853 symbol, as a numeric expression (which may involve labels, and even
5854 forward references) after the type specifier. Like this:
5856 \c global hashtable:data (hashtable.end - hashtable)
5859 \c db this,that,theother ; some data here
5862 This makes NASM automatically calculate the length of the table and
5863 place that information into the \c{ELF} symbol table.
5865 Declaring the type and size of global symbols is necessary when
5866 writing shared library code. For more information, see
5870 \S{elfcomm} \c{elf} Extensions to the \c{COMMON} Directive
5871 \I{COMMON, elf extensions to}
5873 \c{ELF} also allows you to specify alignment requirements \I{common
5874 variables, alignment in elf}\I{alignment, of elf common variables}on
5875 common variables. This is done by putting a number (which must be a
5876 power of two) after the name and size of the common variable,
5877 separated (as usual) by a colon. For example, an array of
5878 doublewords would benefit from 4-byte alignment:
5880 \c common dwordarray 128:4
5882 This declares the total size of the array to be 128 bytes, and
5883 requires that it be aligned on a 4-byte boundary.
5886 \S{elf16} 16-bit code and ELF
5887 \I{ELF, 16-bit code and}
5889 The \c{ELF32} specification doesn't provide relocations for 8- and
5890 16-bit values, but the GNU \c{ld} linker adds these as an extension.
5891 NASM can generate GNU-compatible relocations, to allow 16-bit code to
5892 be linked as ELF using GNU \c{ld}. If NASM is used with the
5893 \c{-w+gnu-elf-extensions} option, a warning is issued when one of
5894 these relocations is generated.
5896 \S{elfdbg} Debug formats and ELF
5897 \I{ELF, Debug formats and}
5899 ELF provides debug information in \c{STABS} and \c{DWARF} formats.
5900 Line number information is generated for all executable sections, but please
5901 note that only the ".text" section is executable by default.
5903 \H{aoutfmt} \i\c{aout}: Linux \I{a.out, Linux version}\I{linux, a.out}\c{a.out} Object Files
5905 The \c{aout} format generates \c{a.out} object files, in the form used
5906 by early Linux systems (current Linux systems use ELF, see
5907 \k{elffmt}.) These differ from other \c{a.out} object files in that
5908 the magic number in the first four bytes of the file is
5909 different; also, some implementations of \c{a.out}, for example
5910 NetBSD's, support position-independent code, which Linux's
5911 implementation does not.
5913 \c{a.out} provides a default output file-name extension of \c{.o}.
5915 \c{a.out} is a very simple object format. It supports no special
5916 directives, no special symbols, no use of \c{SEG} or \c{WRT}, and no
5917 extensions to any standard directives. It supports only the three
5918 \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}.
5921 \H{aoutfmt} \i\c{aoutb}: \i{NetBSD}/\i{FreeBSD}/\i{OpenBSD}
5922 \I{a.out, BSD version}\c{a.out} Object Files
5924 The \c{aoutb} format generates \c{a.out} object files, in the form
5925 used by the various free \c{BSD Unix} clones, \c{NetBSD}, \c{FreeBSD}
5926 and \c{OpenBSD}. For simple object files, this object format is exactly
5927 the same as \c{aout} except for the magic number in the first four bytes
5928 of the file. However, the \c{aoutb} format supports
5929 \I{PIC}\i{position-independent code} in the same way as the \c{elf}
5930 format, so you can use it to write \c{BSD} \i{shared libraries}.
5932 \c{aoutb} provides a default output file-name extension of \c{.o}.
5934 \c{aoutb} supports no special directives, no special symbols, and
5935 only the three \i{standard section names} \i\c{.text}, \i\c{.data}
5936 and \i\c{.bss}. However, it also supports the same use of \i\c{WRT} as
5937 \c{elf} does, to provide position-independent code relocation types.
5938 See \k{elfwrt} for full documentation of this feature.
5940 \c{aoutb} also supports the same extensions to the \c{GLOBAL}
5941 directive as \c{elf} does: see \k{elfglob} for documentation of
5945 \H{as86fmt} \c{as86}: \i{Minix}/Linux\I{linux, as86} \i\c{as86} Object Files
5947 The Minix/Linux 16-bit assembler \c{as86} has its own non-standard
5948 object file format. Although its companion linker \i\c{ld86} produces
5949 something close to ordinary \c{a.out} binaries as output, the object
5950 file format used to communicate between \c{as86} and \c{ld86} is not
5953 NASM supports this format, just in case it is useful, as \c{as86}.
5954 \c{as86} provides a default output file-name extension of \c{.o}.
5956 \c{as86} is a very simple object format (from the NASM user's point
5957 of view). It supports no special directives, no use of \c{SEG} or \c{WRT},
5958 and no extensions to any standard directives. It supports only the three
5959 \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}. The
5960 only special symbol supported is \c{..start}.
5963 \H{rdffmt} \I{RDOFF}\i\c{rdf}: \i{Relocatable Dynamic Object File
5966 The \c{rdf} output format produces \c{RDOFF} object files. \c{RDOFF}
5967 (Relocatable Dynamic Object File Format) is a home-grown object-file
5968 format, designed alongside NASM itself and reflecting in its file
5969 format the internal structure of the assembler.
5971 \c{RDOFF} is not used by any well-known operating systems. Those
5972 writing their own systems, however, may well wish to use \c{RDOFF}
5973 as their object format, on the grounds that it is designed primarily
5974 for simplicity and contains very little file-header bureaucracy.
5976 The Unix NASM archive, and the DOS archive which includes sources,
5977 both contain an \I{rdoff subdirectory}\c{rdoff} subdirectory holding
5978 a set of RDOFF utilities: an RDF linker, an \c{RDF} static-library
5979 manager, an RDF file dump utility, and a program which will load and
5980 execute an RDF executable under Linux.
5982 \c{rdf} supports only the \i{standard section names} \i\c{.text},
5983 \i\c{.data} and \i\c{.bss}.
5986 \S{rdflib} Requiring a Library: The \i\c{LIBRARY} Directive
5988 \c{RDOFF} contains a mechanism for an object file to demand a given
5989 library to be linked to the module, either at load time or run time.
5990 This is done by the \c{LIBRARY} directive, which takes one argument
5991 which is the name of the module:
5993 \c library mylib.rdl
5996 \S{rdfmod} Specifying a Module Name: The \i\c{MODULE} Directive
5998 Special \c{RDOFF} header record is used to store the name of the module.
5999 It can be used, for example, by run-time loader to perform dynamic
6000 linking. \c{MODULE} directive takes one argument which is the name
6005 Note that when you statically link modules and tell linker to strip
6006 the symbols from output file, all module names will be stripped too.
6007 To avoid it, you should start module names with \I{$, prefix}\c{$}, like:
6009 \c module $kernel.core
6012 \S{rdfglob} \c{rdf} Extensions to the \c{GLOBAL} Directive\I{GLOBAL,
6015 \c{RDOFF} global symbols can contain additional information needed by
6016 the static linker. You can mark a global symbol as exported, thus
6017 telling the linker do not strip it from target executable or library
6018 file. Like in \c{ELF}, you can also specify whether an exported symbol
6019 is a procedure (function) or data object.
6021 Suffixing the name with a colon and the word \i\c{export} you make the
6024 \c global sys_open:export
6026 To specify that exported symbol is a procedure (function), you add the
6027 word \i\c{proc} or \i\c{function} after declaration:
6029 \c global sys_open:export proc
6031 Similarly, to specify exported data object, add the word \i\c{data}
6032 or \i\c{object} to the directive:
6034 \c global kernel_ticks:export data
6037 \S{rdfimpt} \c{rdf} Extensions to the \c{EXTERN} Directive\I{EXTERN,
6040 By default the \c{EXTERN} directive in \c{RDOFF} declares a "pure external"
6041 symbol (i.e. the static linker will complain if such a symbol is not resolved).
6042 To declare an "imported" symbol, which must be resolved later during a dynamic
6043 linking phase, \c{RDOFF} offers an additional \c{import} modifier. As in
6044 \c{GLOBAL}, you can also specify whether an imported symbol is a procedure
6045 (function) or data object. For example:
6048 \c extern _open:import
6049 \c extern _printf:import proc
6050 \c extern _errno:import data
6052 Here the directive \c{LIBRARY} is also included, which gives the dynamic linker
6053 a hint as to where to find requested symbols.
6056 \H{dbgfmt} \i\c{dbg}: Debugging Format
6058 The \c{dbg} output format is not built into NASM in the default
6059 configuration. If you are building your own NASM executable from the
6060 sources, you can define \i\c{OF_DBG} in \c{output/outform.h} or on the
6061 compiler command line, and obtain the \c{dbg} output format.
6063 The \c{dbg} format does not output an object file as such; instead,
6064 it outputs a text file which contains a complete list of all the
6065 transactions between the main body of NASM and the output-format
6066 back end module. It is primarily intended to aid people who want to
6067 write their own output drivers, so that they can get a clearer idea
6068 of the various requests the main program makes of the output driver,
6069 and in what order they happen.
6071 For simple files, one can easily use the \c{dbg} format like this:
6073 \c nasm -f dbg filename.asm
6075 which will generate a diagnostic file called \c{filename.dbg}.
6076 However, this will not work well on files which were designed for a
6077 different object format, because each object format defines its own
6078 macros (usually user-level forms of directives), and those macros
6079 will not be defined in the \c{dbg} format. Therefore it can be
6080 useful to run NASM twice, in order to do the preprocessing with the
6081 native object format selected:
6083 \c nasm -e -f rdf -o rdfprog.i rdfprog.asm
6084 \c nasm -a -f dbg rdfprog.i
6086 This preprocesses \c{rdfprog.asm} into \c{rdfprog.i}, keeping the
6087 \c{rdf} object format selected in order to make sure RDF special
6088 directives are converted into primitive form correctly. Then the
6089 preprocessed source is fed through the \c{dbg} format to generate
6090 the final diagnostic output.
6092 This workaround will still typically not work for programs intended
6093 for \c{obj} format, because the \c{obj} \c{SEGMENT} and \c{GROUP}
6094 directives have side effects of defining the segment and group names
6095 as symbols; \c{dbg} will not do this, so the program will not
6096 assemble. You will have to work around that by defining the symbols
6097 yourself (using \c{EXTERN}, for example) if you really need to get a
6098 \c{dbg} trace of an \c{obj}-specific source file.
6100 \c{dbg} accepts any section name and any directives at all, and logs
6101 them all to its output file.
6104 \C{16bit} Writing 16-bit Code (DOS, Windows 3/3.1)
6106 This chapter attempts to cover some of the common issues encountered
6107 when writing 16-bit code to run under \c{MS-DOS} or \c{Windows 3.x}. It
6108 covers how to link programs to produce \c{.EXE} or \c{.COM} files,
6109 how to write \c{.SYS} device drivers, and how to interface assembly
6110 language code with 16-bit C compilers and with Borland Pascal.
6113 \H{exefiles} Producing \i\c{.EXE} Files
6115 Any large program written under DOS needs to be built as a \c{.EXE}
6116 file: only \c{.EXE} files have the necessary internal structure
6117 required to span more than one 64K segment. \i{Windows} programs,
6118 also, have to be built as \c{.EXE} files, since Windows does not
6119 support the \c{.COM} format.
6121 In general, you generate \c{.EXE} files by using the \c{obj} output
6122 format to produce one or more \i\c{.OBJ} files, and then linking
6123 them together using a linker. However, NASM also supports the direct
6124 generation of simple DOS \c{.EXE} files using the \c{bin} output
6125 format (by using \c{DB} and \c{DW} to construct the \c{.EXE} file
6126 header), and a macro package is supplied to do this. Thanks to
6127 Yann Guidon for contributing the code for this.
6129 NASM may also support \c{.EXE} natively as another output format in
6133 \S{objexe} Using the \c{obj} Format To Generate \c{.EXE} Files
6135 This section describes the usual method of generating \c{.EXE} files
6136 by linking \c{.OBJ} files together.
6138 Most 16-bit programming language packages come with a suitable
6139 linker; if you have none of these, there is a free linker called
6140 \i{VAL}\I{linker, free}, available in \c{LZH} archive format from
6141 \W{ftp://x2ftp.oulu.fi/pub/msdos/programming/lang/}\i\c{x2ftp.oulu.fi}.
6142 An LZH archiver can be found at
6143 \W{ftp://ftp.simtel.net/pub/simtelnet/msdos/arcers}\i\c{ftp.simtel.net}.
6144 There is another `free' linker (though this one doesn't come with
6145 sources) called \i{FREELINK}, available from
6146 \W{http://www.pcorner.com/tpc/old/3-101.html}\i\c{www.pcorner.com}.
6147 A third, \i\c{djlink}, written by DJ Delorie, is available at
6148 \W{http://www.delorie.com/djgpp/16bit/djlink/}\i\c{www.delorie.com}.
6149 A fourth linker, \i\c{ALINK}, written by Anthony A.J. Williams, is
6150 available at \W{http://alink.sourceforge.net}\i\c{alink.sourceforge.net}.
6152 When linking several \c{.OBJ} files into a \c{.EXE} file, you should
6153 ensure that exactly one of them has a start point defined (using the
6154 \I{program entry point}\i\c{..start} special symbol defined by the
6155 \c{obj} format: see \k{dotdotstart}). If no module defines a start
6156 point, the linker will not know what value to give the entry-point
6157 field in the output file header; if more than one defines a start
6158 point, the linker will not know \e{which} value to use.
6160 An example of a NASM source file which can be assembled to a
6161 \c{.OBJ} file and linked on its own to a \c{.EXE} is given here. It
6162 demonstrates the basic principles of defining a stack, initialising
6163 the segment registers, and declaring a start point. This file is
6164 also provided in the \I{test subdirectory}\c{test} subdirectory of
6165 the NASM archives, under the name \c{objexe.asm}.
6176 This initial piece of code sets up \c{DS} to point to the data
6177 segment, and initializes \c{SS} and \c{SP} to point to the top of
6178 the provided stack. Notice that interrupts are implicitly disabled
6179 for one instruction after a move into \c{SS}, precisely for this
6180 situation, so that there's no chance of an interrupt occurring
6181 between the loads of \c{SS} and \c{SP} and not having a stack to
6184 Note also that the special symbol \c{..start} is defined at the
6185 beginning of this code, which means that will be the entry point
6186 into the resulting executable file.
6192 The above is the main program: load \c{DS:DX} with a pointer to the
6193 greeting message (\c{hello} is implicitly relative to the segment
6194 \c{data}, which was loaded into \c{DS} in the setup code, so the
6195 full pointer is valid), and call the DOS print-string function.
6200 This terminates the program using another DOS system call.
6204 \c hello: db 'hello, world', 13, 10, '$'
6206 The data segment contains the string we want to display.
6208 \c segment stack stack
6212 The above code declares a stack segment containing 64 bytes of
6213 uninitialized stack space, and points \c{stacktop} at the top of it.
6214 The directive \c{segment stack stack} defines a segment \e{called}
6215 \c{stack}, and also of \e{type} \c{STACK}. The latter is not
6216 necessary to the correct running of the program, but linkers are
6217 likely to issue warnings or errors if your program has no segment of
6220 The above file, when assembled into a \c{.OBJ} file, will link on
6221 its own to a valid \c{.EXE} file, which when run will print `hello,
6222 world' and then exit.
6225 \S{binexe} Using the \c{bin} Format To Generate \c{.EXE} Files
6227 The \c{.EXE} file format is simple enough that it's possible to
6228 build a \c{.EXE} file by writing a pure-binary program and sticking
6229 a 32-byte header on the front. This header is simple enough that it
6230 can be generated using \c{DB} and \c{DW} commands by NASM itself, so
6231 that you can use the \c{bin} output format to directly generate
6234 Included in the NASM archives, in the \I{misc subdirectory}\c{misc}
6235 subdirectory, is a file \i\c{exebin.mac} of macros. It defines three
6236 macros: \i\c{EXE_begin}, \i\c{EXE_stack} and \i\c{EXE_end}.
6238 To produce a \c{.EXE} file using this method, you should start by
6239 using \c{%include} to load the \c{exebin.mac} macro package into
6240 your source file. You should then issue the \c{EXE_begin} macro call
6241 (which takes no arguments) to generate the file header data. Then
6242 write code as normal for the \c{bin} format - you can use all three
6243 standard sections \c{.text}, \c{.data} and \c{.bss}. At the end of
6244 the file you should call the \c{EXE_end} macro (again, no arguments),
6245 which defines some symbols to mark section sizes, and these symbols
6246 are referred to in the header code generated by \c{EXE_begin}.
6248 In this model, the code you end up writing starts at \c{0x100}, just
6249 like a \c{.COM} file - in fact, if you strip off the 32-byte header
6250 from the resulting \c{.EXE} file, you will have a valid \c{.COM}
6251 program. All the segment bases are the same, so you are limited to a
6252 64K program, again just like a \c{.COM} file. Note that an \c{ORG}
6253 directive is issued by the \c{EXE_begin} macro, so you should not
6254 explicitly issue one of your own.
6256 You can't directly refer to your segment base value, unfortunately,
6257 since this would require a relocation in the header, and things
6258 would get a lot more complicated. So you should get your segment
6259 base by copying it out of \c{CS} instead.
6261 On entry to your \c{.EXE} file, \c{SS:SP} are already set up to
6262 point to the top of a 2Kb stack. You can adjust the default stack
6263 size of 2Kb by calling the \c{EXE_stack} macro. For example, to
6264 change the stack size of your program to 64 bytes, you would call
6267 A sample program which generates a \c{.EXE} file in this way is
6268 given in the \c{test} subdirectory of the NASM archive, as
6272 \H{comfiles} Producing \i\c{.COM} Files
6274 While large DOS programs must be written as \c{.EXE} files, small
6275 ones are often better written as \c{.COM} files. \c{.COM} files are
6276 pure binary, and therefore most easily produced using the \c{bin}
6280 \S{combinfmt} Using the \c{bin} Format To Generate \c{.COM} Files
6282 \c{.COM} files expect to be loaded at offset \c{100h} into their
6283 segment (though the segment may change). Execution then begins at
6284 \I\c{ORG}\c{100h}, i.e. right at the start of the program. So to
6285 write a \c{.COM} program, you would create a source file looking
6293 \c ; put your code here
6297 \c ; put data items here
6301 \c ; put uninitialized data here
6303 The \c{bin} format puts the \c{.text} section first in the file, so
6304 you can declare data or BSS items before beginning to write code if
6305 you want to and the code will still end up at the front of the file
6308 The BSS (uninitialized data) section does not take up space in the
6309 \c{.COM} file itself: instead, addresses of BSS items are resolved
6310 to point at space beyond the end of the file, on the grounds that
6311 this will be free memory when the program is run. Therefore you
6312 should not rely on your BSS being initialized to all zeros when you
6315 To assemble the above program, you should use a command line like
6317 \c nasm myprog.asm -fbin -o myprog.com
6319 The \c{bin} format would produce a file called \c{myprog} if no
6320 explicit output file name were specified, so you have to override it
6321 and give the desired file name.
6324 \S{comobjfmt} Using the \c{obj} Format To Generate \c{.COM} Files
6326 If you are writing a \c{.COM} program as more than one module, you
6327 may wish to assemble several \c{.OBJ} files and link them together
6328 into a \c{.COM} program. You can do this, provided you have a linker
6329 capable of outputting \c{.COM} files directly (\i{TLINK} does this),
6330 or alternatively a converter program such as \i\c{EXE2BIN} to
6331 transform the \c{.EXE} file output from the linker into a \c{.COM}
6334 If you do this, you need to take care of several things:
6336 \b The first object file containing code should start its code
6337 segment with a line like \c{RESB 100h}. This is to ensure that the
6338 code begins at offset \c{100h} relative to the beginning of the code
6339 segment, so that the linker or converter program does not have to
6340 adjust address references within the file when generating the
6341 \c{.COM} file. Other assemblers use an \i\c{ORG} directive for this
6342 purpose, but \c{ORG} in NASM is a format-specific directive to the
6343 \c{bin} output format, and does not mean the same thing as it does
6344 in MASM-compatible assemblers.
6346 \b You don't need to define a stack segment.
6348 \b All your segments should be in the same group, so that every time
6349 your code or data references a symbol offset, all offsets are
6350 relative to the same segment base. This is because, when a \c{.COM}
6351 file is loaded, all the segment registers contain the same value.
6354 \H{sysfiles} Producing \i\c{.SYS} Files
6356 \i{MS-DOS device drivers} - \c{.SYS} files - are pure binary files,
6357 similar to \c{.COM} files, except that they start at origin zero
6358 rather than \c{100h}. Therefore, if you are writing a device driver
6359 using the \c{bin} format, you do not need the \c{ORG} directive,
6360 since the default origin for \c{bin} is zero. Similarly, if you are
6361 using \c{obj}, you do not need the \c{RESB 100h} at the start of
6364 \c{.SYS} files start with a header structure, containing pointers to
6365 the various routines inside the driver which do the work. This
6366 structure should be defined at the start of the code segment, even
6367 though it is not actually code.
6369 For more information on the format of \c{.SYS} files, and the data
6370 which has to go in the header structure, a list of books is given in
6371 the Frequently Asked Questions list for the newsgroup
6372 \W{news:comp.os.msdos.programmer}\i\c{comp.os.msdos.programmer}.
6375 \H{16c} Interfacing to 16-bit C Programs
6377 This section covers the basics of writing assembly routines that
6378 call, or are called from, C programs. To do this, you would
6379 typically write an assembly module as a \c{.OBJ} file, and link it
6380 with your C modules to produce a \i{mixed-language program}.
6383 \S{16cunder} External Symbol Names
6385 \I{C symbol names}\I{underscore, in C symbols}C compilers have the
6386 convention that the names of all global symbols (functions or data)
6387 they define are formed by prefixing an underscore to the name as it
6388 appears in the C program. So, for example, the function a C
6389 programmer thinks of as \c{printf} appears to an assembly language
6390 programmer as \c{_printf}. This means that in your assembly
6391 programs, you can define symbols without a leading underscore, and
6392 not have to worry about name clashes with C symbols.
6394 If you find the underscores inconvenient, you can define macros to
6395 replace the \c{GLOBAL} and \c{EXTERN} directives as follows:
6411 (These forms of the macros only take one argument at a time; a
6412 \c{%rep} construct could solve this.)
6414 If you then declare an external like this:
6418 then the macro will expand it as
6421 \c %define printf _printf
6423 Thereafter, you can reference \c{printf} as if it was a symbol, and
6424 the preprocessor will put the leading underscore on where necessary.
6426 The \c{cglobal} macro works similarly. You must use \c{cglobal}
6427 before defining the symbol in question, but you would have had to do
6428 that anyway if you used \c{GLOBAL}.
6430 Also see \k{opt-pfix}.
6432 \S{16cmodels} \i{Memory Models}
6434 NASM contains no mechanism to support the various C memory models
6435 directly; you have to keep track yourself of which one you are
6436 writing for. This means you have to keep track of the following
6439 \b In models using a single code segment (tiny, small and compact),
6440 functions are near. This means that function pointers, when stored
6441 in data segments or pushed on the stack as function arguments, are
6442 16 bits long and contain only an offset field (the \c{CS} register
6443 never changes its value, and always gives the segment part of the
6444 full function address), and that functions are called using ordinary
6445 near \c{CALL} instructions and return using \c{RETN} (which, in
6446 NASM, is synonymous with \c{RET} anyway). This means both that you
6447 should write your own routines to return with \c{RETN}, and that you
6448 should call external C routines with near \c{CALL} instructions.
6450 \b In models using more than one code segment (medium, large and
6451 huge), functions are far. This means that function pointers are 32
6452 bits long (consisting of a 16-bit offset followed by a 16-bit
6453 segment), and that functions are called using \c{CALL FAR} (or
6454 \c{CALL seg:offset}) and return using \c{RETF}. Again, you should
6455 therefore write your own routines to return with \c{RETF} and use
6456 \c{CALL FAR} to call external routines.
6458 \b In models using a single data segment (tiny, small and medium),
6459 data pointers are 16 bits long, containing only an offset field (the
6460 \c{DS} register doesn't change its value, and always gives the
6461 segment part of the full data item address).
6463 \b In models using more than one data segment (compact, large and
6464 huge), data pointers are 32 bits long, consisting of a 16-bit offset
6465 followed by a 16-bit segment. You should still be careful not to
6466 modify \c{DS} in your routines without restoring it afterwards, but
6467 \c{ES} is free for you to use to access the contents of 32-bit data
6468 pointers you are passed.
6470 \b The huge memory model allows single data items to exceed 64K in
6471 size. In all other memory models, you can access the whole of a data
6472 item just by doing arithmetic on the offset field of the pointer you
6473 are given, whether a segment field is present or not; in huge model,
6474 you have to be more careful of your pointer arithmetic.
6476 \b In most memory models, there is a \e{default} data segment, whose
6477 segment address is kept in \c{DS} throughout the program. This data
6478 segment is typically the same segment as the stack, kept in \c{SS},
6479 so that functions' local variables (which are stored on the stack)
6480 and global data items can both be accessed easily without changing
6481 \c{DS}. Particularly large data items are typically stored in other
6482 segments. However, some memory models (though not the standard
6483 ones, usually) allow the assumption that \c{SS} and \c{DS} hold the
6484 same value to be removed. Be careful about functions' local
6485 variables in this latter case.
6487 In models with a single code segment, the segment is called
6488 \i\c{_TEXT}, so your code segment must also go by this name in order
6489 to be linked into the same place as the main code segment. In models
6490 with a single data segment, or with a default data segment, it is
6494 \S{16cfunc} Function Definitions and Function Calls
6496 \I{functions, C calling convention}The \i{C calling convention} in
6497 16-bit programs is as follows. In the following description, the
6498 words \e{caller} and \e{callee} are used to denote the function
6499 doing the calling and the function which gets called.
6501 \b The caller pushes the function's parameters on the stack, one
6502 after another, in reverse order (right to left, so that the first
6503 argument specified to the function is pushed last).
6505 \b The caller then executes a \c{CALL} instruction to pass control
6506 to the callee. This \c{CALL} is either near or far depending on the
6509 \b The callee receives control, and typically (although this is not
6510 actually necessary, in functions which do not need to access their
6511 parameters) starts by saving the value of \c{SP} in \c{BP} so as to
6512 be able to use \c{BP} as a base pointer to find its parameters on
6513 the stack. However, the caller was probably doing this too, so part
6514 of the calling convention states that \c{BP} must be preserved by
6515 any C function. Hence the callee, if it is going to set up \c{BP} as
6516 a \i\e{frame pointer}, must push the previous value first.
6518 \b The callee may then access its parameters relative to \c{BP}.
6519 The word at \c{[BP]} holds the previous value of \c{BP} as it was
6520 pushed; the next word, at \c{[BP+2]}, holds the offset part of the
6521 return address, pushed implicitly by \c{CALL}. In a small-model
6522 (near) function, the parameters start after that, at \c{[BP+4]}; in
6523 a large-model (far) function, the segment part of the return address
6524 lives at \c{[BP+4]}, and the parameters begin at \c{[BP+6]}. The
6525 leftmost parameter of the function, since it was pushed last, is
6526 accessible at this offset from \c{BP}; the others follow, at
6527 successively greater offsets. Thus, in a function such as \c{printf}
6528 which takes a variable number of parameters, the pushing of the
6529 parameters in reverse order means that the function knows where to
6530 find its first parameter, which tells it the number and type of the
6533 \b The callee may also wish to decrease \c{SP} further, so as to
6534 allocate space on the stack for local variables, which will then be
6535 accessible at negative offsets from \c{BP}.
6537 \b The callee, if it wishes to return a value to the caller, should
6538 leave the value in \c{AL}, \c{AX} or \c{DX:AX} depending on the size
6539 of the value. Floating-point results are sometimes (depending on the
6540 compiler) returned in \c{ST0}.
6542 \b Once the callee has finished processing, it restores \c{SP} from
6543 \c{BP} if it had allocated local stack space, then pops the previous
6544 value of \c{BP}, and returns via \c{RETN} or \c{RETF} depending on
6547 \b When the caller regains control from the callee, the function
6548 parameters are still on the stack, so it typically adds an immediate
6549 constant to \c{SP} to remove them (instead of executing a number of
6550 slow \c{POP} instructions). Thus, if a function is accidentally
6551 called with the wrong number of parameters due to a prototype
6552 mismatch, the stack will still be returned to a sensible state since
6553 the caller, which \e{knows} how many parameters it pushed, does the
6556 It is instructive to compare this calling convention with that for
6557 Pascal programs (described in \k{16bpfunc}). Pascal has a simpler
6558 convention, since no functions have variable numbers of parameters.
6559 Therefore the callee knows how many parameters it should have been
6560 passed, and is able to deallocate them from the stack itself by
6561 passing an immediate argument to the \c{RET} or \c{RETF}
6562 instruction, so the caller does not have to do it. Also, the
6563 parameters are pushed in left-to-right order, not right-to-left,
6564 which means that a compiler can give better guarantees about
6565 sequence points without performance suffering.
6567 Thus, you would define a function in C style in the following way.
6568 The following example is for small model:
6575 \c sub sp,0x40 ; 64 bytes of local stack space
6576 \c mov bx,[bp+4] ; first parameter to function
6580 \c mov sp,bp ; undo "sub sp,0x40" above
6584 For a large-model function, you would replace \c{RET} by \c{RETF},
6585 and look for the first parameter at \c{[BP+6]} instead of
6586 \c{[BP+4]}. Of course, if one of the parameters is a pointer, then
6587 the offsets of \e{subsequent} parameters will change depending on
6588 the memory model as well: far pointers take up four bytes on the
6589 stack when passed as a parameter, whereas near pointers take up two.
6591 At the other end of the process, to call a C function from your
6592 assembly code, you would do something like this:
6596 \c ; and then, further down...
6598 \c push word [myint] ; one of my integer variables
6599 \c push word mystring ; pointer into my data segment
6601 \c add sp,byte 4 ; `byte' saves space
6603 \c ; then those data items...
6608 \c mystring db 'This number -> %d <- should be 1234',10,0
6610 This piece of code is the small-model assembly equivalent of the C
6613 \c int myint = 1234;
6614 \c printf("This number -> %d <- should be 1234\n", myint);
6616 In large model, the function-call code might look more like this. In
6617 this example, it is assumed that \c{DS} already holds the segment
6618 base of the segment \c{_DATA}. If not, you would have to initialize
6621 \c push word [myint]
6622 \c push word seg mystring ; Now push the segment, and...
6623 \c push word mystring ; ... offset of "mystring"
6627 The integer value still takes up one word on the stack, since large
6628 model does not affect the size of the \c{int} data type. The first
6629 argument (pushed last) to \c{printf}, however, is a data pointer,
6630 and therefore has to contain a segment and offset part. The segment
6631 should be stored second in memory, and therefore must be pushed
6632 first. (Of course, \c{PUSH DS} would have been a shorter instruction
6633 than \c{PUSH WORD SEG mystring}, if \c{DS} was set up as the above
6634 example assumed.) Then the actual call becomes a far call, since
6635 functions expect far calls in large model; and \c{SP} has to be
6636 increased by 6 rather than 4 afterwards to make up for the extra
6640 \S{16cdata} Accessing Data Items
6642 To get at the contents of C variables, or to declare variables which
6643 C can access, you need only declare the names as \c{GLOBAL} or
6644 \c{EXTERN}. (Again, the names require leading underscores, as stated
6645 in \k{16cunder}.) Thus, a C variable declared as \c{int i} can be
6646 accessed from assembler as
6652 And to declare your own integer variable which C programs can access
6653 as \c{extern int j}, you do this (making sure you are assembling in
6654 the \c{_DATA} segment, if necessary):
6660 To access a C array, you need to know the size of the components of
6661 the array. For example, \c{int} variables are two bytes long, so if
6662 a C program declares an array as \c{int a[10]}, you can access
6663 \c{a[3]} by coding \c{mov ax,[_a+6]}. (The byte offset 6 is obtained
6664 by multiplying the desired array index, 3, by the size of the array
6665 element, 2.) The sizes of the C base types in 16-bit compilers are:
6666 1 for \c{char}, 2 for \c{short} and \c{int}, 4 for \c{long} and
6667 \c{float}, and 8 for \c{double}.
6669 To access a C \i{data structure}, you need to know the offset from
6670 the base of the structure to the field you are interested in. You
6671 can either do this by converting the C structure definition into a
6672 NASM structure definition (using \i\c{STRUC}), or by calculating the
6673 one offset and using just that.
6675 To do either of these, you should read your C compiler's manual to
6676 find out how it organizes data structures. NASM gives no special
6677 alignment to structure members in its own \c{STRUC} macro, so you
6678 have to specify alignment yourself if the C compiler generates it.
6679 Typically, you might find that a structure like
6686 might be four bytes long rather than three, since the \c{int} field
6687 would be aligned to a two-byte boundary. However, this sort of
6688 feature tends to be a configurable option in the C compiler, either
6689 using command-line options or \c{#pragma} lines, so you have to find
6690 out how your own compiler does it.
6693 \S{16cmacro} \i\c{c16.mac}: Helper Macros for the 16-bit C Interface
6695 Included in the NASM archives, in the \I{misc subdirectory}\c{misc}
6696 directory, is a file \c{c16.mac} of macros. It defines three macros:
6697 \i\c{proc}, \i\c{arg} and \i\c{endproc}. These are intended to be
6698 used for C-style procedure definitions, and they automate a lot of
6699 the work involved in keeping track of the calling convention.
6701 (An alternative, TASM compatible form of \c{arg} is also now built
6702 into NASM's preprocessor. See \k{stackrel} for details.)
6704 An example of an assembly function using the macro set is given
6711 \c mov ax,[bp + %$i]
6712 \c mov bx,[bp + %$j]
6717 This defines \c{_nearproc} to be a procedure taking two arguments,
6718 the first (\c{i}) an integer and the second (\c{j}) a pointer to an
6719 integer. It returns \c{i + *j}.
6721 Note that the \c{arg} macro has an \c{EQU} as the first line of its
6722 expansion, and since the label before the macro call gets prepended
6723 to the first line of the expanded macro, the \c{EQU} works, defining
6724 \c{%$i} to be an offset from \c{BP}. A context-local variable is
6725 used, local to the context pushed by the \c{proc} macro and popped
6726 by the \c{endproc} macro, so that the same argument name can be used
6727 in later procedures. Of course, you don't \e{have} to do that.
6729 The macro set produces code for near functions (tiny, small and
6730 compact-model code) by default. You can have it generate far
6731 functions (medium, large and huge-model code) by means of coding
6732 \I\c{FARCODE}\c{%define FARCODE}. This changes the kind of return
6733 instruction generated by \c{endproc}, and also changes the starting
6734 point for the argument offsets. The macro set contains no intrinsic
6735 dependency on whether data pointers are far or not.
6737 \c{arg} can take an optional parameter, giving the size of the
6738 argument. If no size is given, 2 is assumed, since it is likely that
6739 many function parameters will be of type \c{int}.
6741 The large-model equivalent of the above function would look like this:
6749 \c mov ax,[bp + %$i]
6750 \c mov bx,[bp + %$j]
6751 \c mov es,[bp + %$j + 2]
6756 This makes use of the argument to the \c{arg} macro to define a
6757 parameter of size 4, because \c{j} is now a far pointer. When we
6758 load from \c{j}, we must load a segment and an offset.
6761 \H{16bp} Interfacing to \i{Borland Pascal} Programs
6763 Interfacing to Borland Pascal programs is similar in concept to
6764 interfacing to 16-bit C programs. The differences are:
6766 \b The leading underscore required for interfacing to C programs is
6767 not required for Pascal.
6769 \b The memory model is always large: functions are far, data
6770 pointers are far, and no data item can be more than 64K long.
6771 (Actually, some functions are near, but only those functions that
6772 are local to a Pascal unit and never called from outside it. All
6773 assembly functions that Pascal calls, and all Pascal functions that
6774 assembly routines are able to call, are far.) However, all static
6775 data declared in a Pascal program goes into the default data
6776 segment, which is the one whose segment address will be in \c{DS}
6777 when control is passed to your assembly code. The only things that
6778 do not live in the default data segment are local variables (they
6779 live in the stack segment) and dynamically allocated variables. All
6780 data \e{pointers}, however, are far.
6782 \b The function calling convention is different - described below.
6784 \b Some data types, such as strings, are stored differently.
6786 \b There are restrictions on the segment names you are allowed to
6787 use - Borland Pascal will ignore code or data declared in a segment
6788 it doesn't like the name of. The restrictions are described below.
6791 \S{16bpfunc} The Pascal Calling Convention
6793 \I{functions, Pascal calling convention}\I{Pascal calling
6794 convention}The 16-bit Pascal calling convention is as follows. In
6795 the following description, the words \e{caller} and \e{callee} are
6796 used to denote the function doing the calling and the function which
6799 \b The caller pushes the function's parameters on the stack, one
6800 after another, in normal order (left to right, so that the first
6801 argument specified to the function is pushed first).
6803 \b The caller then executes a far \c{CALL} instruction to pass
6804 control to the callee.
6806 \b The callee receives control, and typically (although this is not
6807 actually necessary, in functions which do not need to access their
6808 parameters) starts by saving the value of \c{SP} in \c{BP} so as to
6809 be able to use \c{BP} as a base pointer to find its parameters on
6810 the stack. However, the caller was probably doing this too, so part
6811 of the calling convention states that \c{BP} must be preserved by
6812 any function. Hence the callee, if it is going to set up \c{BP} as a
6813 \i{frame pointer}, must push the previous value first.
6815 \b The callee may then access its parameters relative to \c{BP}.
6816 The word at \c{[BP]} holds the previous value of \c{BP} as it was
6817 pushed. The next word, at \c{[BP+2]}, holds the offset part of the
6818 return address, and the next one at \c{[BP+4]} the segment part. The
6819 parameters begin at \c{[BP+6]}. The rightmost parameter of the
6820 function, since it was pushed last, is accessible at this offset
6821 from \c{BP}; the others follow, at successively greater offsets.
6823 \b The callee may also wish to decrease \c{SP} further, so as to
6824 allocate space on the stack for local variables, which will then be
6825 accessible at negative offsets from \c{BP}.
6827 \b The callee, if it wishes to return a value to the caller, should
6828 leave the value in \c{AL}, \c{AX} or \c{DX:AX} depending on the size
6829 of the value. Floating-point results are returned in \c{ST0}.
6830 Results of type \c{Real} (Borland's own custom floating-point data
6831 type, not handled directly by the FPU) are returned in \c{DX:BX:AX}.
6832 To return a result of type \c{String}, the caller pushes a pointer
6833 to a temporary string before pushing the parameters, and the callee
6834 places the returned string value at that location. The pointer is
6835 not a parameter, and should not be removed from the stack by the
6836 \c{RETF} instruction.
6838 \b Once the callee has finished processing, it restores \c{SP} from
6839 \c{BP} if it had allocated local stack space, then pops the previous
6840 value of \c{BP}, and returns via \c{RETF}. It uses the form of
6841 \c{RETF} with an immediate parameter, giving the number of bytes
6842 taken up by the parameters on the stack. This causes the parameters
6843 to be removed from the stack as a side effect of the return
6846 \b When the caller regains control from the callee, the function
6847 parameters have already been removed from the stack, so it needs to
6850 Thus, you would define a function in Pascal style, taking two
6851 \c{Integer}-type parameters, in the following way:
6857 \c sub sp,0x40 ; 64 bytes of local stack space
6858 \c mov bx,[bp+8] ; first parameter to function
6859 \c mov bx,[bp+6] ; second parameter to function
6863 \c mov sp,bp ; undo "sub sp,0x40" above
6865 \c retf 4 ; total size of params is 4
6867 At the other end of the process, to call a Pascal function from your
6868 assembly code, you would do something like this:
6872 \c ; and then, further down...
6874 \c push word seg mystring ; Now push the segment, and...
6875 \c push word mystring ; ... offset of "mystring"
6876 \c push word [myint] ; one of my variables
6877 \c call far SomeFunc
6879 This is equivalent to the Pascal code
6881 \c procedure SomeFunc(String: PChar; Int: Integer);
6882 \c SomeFunc(@mystring, myint);
6885 \S{16bpseg} Borland Pascal \I{segment names, Borland Pascal}Segment
6888 Since Borland Pascal's internal unit file format is completely
6889 different from \c{OBJ}, it only makes a very sketchy job of actually
6890 reading and understanding the various information contained in a
6891 real \c{OBJ} file when it links that in. Therefore an object file
6892 intended to be linked to a Pascal program must obey a number of
6895 \b Procedures and functions must be in a segment whose name is
6896 either \c{CODE}, \c{CSEG}, or something ending in \c{_TEXT}.
6898 \b initialized data must be in a segment whose name is either
6899 \c{CONST} or something ending in \c{_DATA}.
6901 \b Uninitialized data must be in a segment whose name is either
6902 \c{DATA}, \c{DSEG}, or something ending in \c{_BSS}.
6904 \b Any other segments in the object file are completely ignored.
6905 \c{GROUP} directives and segment attributes are also ignored.
6908 \S{16bpmacro} Using \i\c{c16.mac} With Pascal Programs
6910 The \c{c16.mac} macro package, described in \k{16cmacro}, can also
6911 be used to simplify writing functions to be called from Pascal
6912 programs, if you code \I\c{PASCAL}\c{%define PASCAL}. This
6913 definition ensures that functions are far (it implies
6914 \i\c{FARCODE}), and also causes procedure return instructions to be
6915 generated with an operand.
6917 Defining \c{PASCAL} does not change the code which calculates the
6918 argument offsets; you must declare your function's arguments in
6919 reverse order. For example:
6927 \c mov ax,[bp + %$i]
6928 \c mov bx,[bp + %$j]
6929 \c mov es,[bp + %$j + 2]
6934 This defines the same routine, conceptually, as the example in
6935 \k{16cmacro}: it defines a function taking two arguments, an integer
6936 and a pointer to an integer, which returns the sum of the integer
6937 and the contents of the pointer. The only difference between this
6938 code and the large-model C version is that \c{PASCAL} is defined
6939 instead of \c{FARCODE}, and that the arguments are declared in
6943 \C{32bit} Writing 32-bit Code (Unix, Win32, DJGPP)
6945 This chapter attempts to cover some of the common issues involved
6946 when writing 32-bit code, to run under \i{Win32} or Unix, or to be
6947 linked with C code generated by a Unix-style C compiler such as
6948 \i{DJGPP}. It covers how to write assembly code to interface with
6949 32-bit C routines, and how to write position-independent code for
6952 Almost all 32-bit code, and in particular all code running under
6953 \c{Win32}, \c{DJGPP} or any of the PC Unix variants, runs in \I{flat
6954 memory model}\e{flat} memory model. This means that the segment registers
6955 and paging have already been set up to give you the same 32-bit 4Gb
6956 address space no matter what segment you work relative to, and that
6957 you should ignore all segment registers completely. When writing
6958 flat-model application code, you never need to use a segment
6959 override or modify any segment register, and the code-section
6960 addresses you pass to \c{CALL} and \c{JMP} live in the same address
6961 space as the data-section addresses you access your variables by and
6962 the stack-section addresses you access local variables and procedure
6963 parameters by. Every address is 32 bits long and contains only an
6967 \H{32c} Interfacing to 32-bit C Programs
6969 A lot of the discussion in \k{16c}, about interfacing to 16-bit C
6970 programs, still applies when working in 32 bits. The absence of
6971 memory models or segmentation worries simplifies things a lot.
6974 \S{32cunder} External Symbol Names
6976 Most 32-bit C compilers share the convention used by 16-bit
6977 compilers, that the names of all global symbols (functions or data)
6978 they define are formed by prefixing an underscore to the name as it
6979 appears in the C program. However, not all of them do: the \c{ELF}
6980 specification states that C symbols do \e{not} have a leading
6981 underscore on their assembly-language names.
6983 The older Linux \c{a.out} C compiler, all \c{Win32} compilers,
6984 \c{DJGPP}, and \c{NetBSD} and \c{FreeBSD}, all use the leading
6985 underscore; for these compilers, the macros \c{cextern} and
6986 \c{cglobal}, as given in \k{16cunder}, will still work. For \c{ELF},
6987 though, the leading underscore should not be used.
6989 See also \k{opt-pfix}.
6991 \S{32cfunc} Function Definitions and Function Calls
6993 \I{functions, C calling convention}The \i{C calling convention}
6994 in 32-bit programs is as follows. In the following description,
6995 the words \e{caller} and \e{callee} are used to denote
6996 the function doing the calling and the function which gets called.
6998 \b The caller pushes the function's parameters on the stack, one
6999 after another, in reverse order (right to left, so that the first
7000 argument specified to the function is pushed last).
7002 \b The caller then executes a near \c{CALL} instruction to pass
7003 control to the callee.
7005 \b The callee receives control, and typically (although this is not
7006 actually necessary, in functions which do not need to access their
7007 parameters) starts by saving the value of \c{ESP} in \c{EBP} so as
7008 to be able to use \c{EBP} as a base pointer to find its parameters
7009 on the stack. However, the caller was probably doing this too, so
7010 part of the calling convention states that \c{EBP} must be preserved
7011 by any C function. Hence the callee, if it is going to set up
7012 \c{EBP} as a \i{frame pointer}, must push the previous value first.
7014 \b The callee may then access its parameters relative to \c{EBP}.
7015 The doubleword at \c{[EBP]} holds the previous value of \c{EBP} as
7016 it was pushed; the next doubleword, at \c{[EBP+4]}, holds the return
7017 address, pushed implicitly by \c{CALL}. The parameters start after
7018 that, at \c{[EBP+8]}. The leftmost parameter of the function, since
7019 it was pushed last, is accessible at this offset from \c{EBP}; the
7020 others follow, at successively greater offsets. Thus, in a function
7021 such as \c{printf} which takes a variable number of parameters, the
7022 pushing of the parameters in reverse order means that the function
7023 knows where to find its first parameter, which tells it the number
7024 and type of the remaining ones.
7026 \b The callee may also wish to decrease \c{ESP} further, so as to
7027 allocate space on the stack for local variables, which will then be
7028 accessible at negative offsets from \c{EBP}.
7030 \b The callee, if it wishes to return a value to the caller, should
7031 leave the value in \c{AL}, \c{AX} or \c{EAX} depending on the size
7032 of the value. Floating-point results are typically returned in
7035 \b Once the callee has finished processing, it restores \c{ESP} from
7036 \c{EBP} if it had allocated local stack space, then pops the previous
7037 value of \c{EBP}, and returns via \c{RET} (equivalently, \c{RETN}).
7039 \b When the caller regains control from the callee, the function
7040 parameters are still on the stack, so it typically adds an immediate
7041 constant to \c{ESP} to remove them (instead of executing a number of
7042 slow \c{POP} instructions). Thus, if a function is accidentally
7043 called with the wrong number of parameters due to a prototype
7044 mismatch, the stack will still be returned to a sensible state since
7045 the caller, which \e{knows} how many parameters it pushed, does the
7048 There is an alternative calling convention used by Win32 programs
7049 for Windows API calls, and also for functions called \e{by} the
7050 Windows API such as window procedures: they follow what Microsoft
7051 calls the \c{__stdcall} convention. This is slightly closer to the
7052 Pascal convention, in that the callee clears the stack by passing a
7053 parameter to the \c{RET} instruction. However, the parameters are
7054 still pushed in right-to-left order.
7056 Thus, you would define a function in C style in the following way:
7063 \c sub esp,0x40 ; 64 bytes of local stack space
7064 \c mov ebx,[ebp+8] ; first parameter to function
7068 \c leave ; mov esp,ebp / pop ebp
7071 At the other end of the process, to call a C function from your
7072 assembly code, you would do something like this:
7076 \c ; and then, further down...
7078 \c push dword [myint] ; one of my integer variables
7079 \c push dword mystring ; pointer into my data segment
7081 \c add esp,byte 8 ; `byte' saves space
7083 \c ; then those data items...
7088 \c mystring db 'This number -> %d <- should be 1234',10,0
7090 This piece of code is the assembly equivalent of the C code
7092 \c int myint = 1234;
7093 \c printf("This number -> %d <- should be 1234\n", myint);
7096 \S{32cdata} Accessing Data Items
7098 To get at the contents of C variables, or to declare variables which
7099 C can access, you need only declare the names as \c{GLOBAL} or
7100 \c{EXTERN}. (Again, the names require leading underscores, as stated
7101 in \k{32cunder}.) Thus, a C variable declared as \c{int i} can be
7102 accessed from assembler as
7107 And to declare your own integer variable which C programs can access
7108 as \c{extern int j}, you do this (making sure you are assembling in
7109 the \c{_DATA} segment, if necessary):
7114 To access a C array, you need to know the size of the components of
7115 the array. For example, \c{int} variables are four bytes long, so if
7116 a C program declares an array as \c{int a[10]}, you can access
7117 \c{a[3]} by coding \c{mov ax,[_a+12]}. (The byte offset 12 is obtained
7118 by multiplying the desired array index, 3, by the size of the array
7119 element, 4.) The sizes of the C base types in 32-bit compilers are:
7120 1 for \c{char}, 2 for \c{short}, 4 for \c{int}, \c{long} and
7121 \c{float}, and 8 for \c{double}. Pointers, being 32-bit addresses,
7122 are also 4 bytes long.
7124 To access a C \i{data structure}, you need to know the offset from
7125 the base of the structure to the field you are interested in. You
7126 can either do this by converting the C structure definition into a
7127 NASM structure definition (using \c{STRUC}), or by calculating the
7128 one offset and using just that.
7130 To do either of these, you should read your C compiler's manual to
7131 find out how it organizes data structures. NASM gives no special
7132 alignment to structure members in its own \i\c{STRUC} macro, so you
7133 have to specify alignment yourself if the C compiler generates it.
7134 Typically, you might find that a structure like
7141 might be eight bytes long rather than five, since the \c{int} field
7142 would be aligned to a four-byte boundary. However, this sort of
7143 feature is sometimes a configurable option in the C compiler, either
7144 using command-line options or \c{#pragma} lines, so you have to find
7145 out how your own compiler does it.
7148 \S{32cmacro} \i\c{c32.mac}: Helper Macros for the 32-bit C Interface
7150 Included in the NASM archives, in the \I{misc directory}\c{misc}
7151 directory, is a file \c{c32.mac} of macros. It defines three macros:
7152 \i\c{proc}, \i\c{arg} and \i\c{endproc}. These are intended to be
7153 used for C-style procedure definitions, and they automate a lot of
7154 the work involved in keeping track of the calling convention.
7156 An example of an assembly function using the macro set is given
7163 \c mov eax,[ebp + %$i]
7164 \c mov ebx,[ebp + %$j]
7169 This defines \c{_proc32} to be a procedure taking two arguments, the
7170 first (\c{i}) an integer and the second (\c{j}) a pointer to an
7171 integer. It returns \c{i + *j}.
7173 Note that the \c{arg} macro has an \c{EQU} as the first line of its
7174 expansion, and since the label before the macro call gets prepended
7175 to the first line of the expanded macro, the \c{EQU} works, defining
7176 \c{%$i} to be an offset from \c{BP}. A context-local variable is
7177 used, local to the context pushed by the \c{proc} macro and popped
7178 by the \c{endproc} macro, so that the same argument name can be used
7179 in later procedures. Of course, you don't \e{have} to do that.
7181 \c{arg} can take an optional parameter, giving the size of the
7182 argument. If no size is given, 4 is assumed, since it is likely that
7183 many function parameters will be of type \c{int} or pointers.
7186 \H{picdll} Writing NetBSD/FreeBSD/OpenBSD and Linux/ELF \i{Shared
7189 \c{ELF} replaced the older \c{a.out} object file format under Linux
7190 because it contains support for \i{position-independent code}
7191 (\i{PIC}), which makes writing shared libraries much easier. NASM
7192 supports the \c{ELF} position-independent code features, so you can
7193 write Linux \c{ELF} shared libraries in NASM.
7195 \i{NetBSD}, and its close cousins \i{FreeBSD} and \i{OpenBSD}, take
7196 a different approach by hacking PIC support into the \c{a.out}
7197 format. NASM supports this as the \i\c{aoutb} output format, so you
7198 can write \i{BSD} shared libraries in NASM too.
7200 The operating system loads a PIC shared library by memory-mapping
7201 the library file at an arbitrarily chosen point in the address space
7202 of the running process. The contents of the library's code section
7203 must therefore not depend on where it is loaded in memory.
7205 Therefore, you cannot get at your variables by writing code like
7208 \c mov eax,[myvar] ; WRONG
7210 Instead, the linker provides an area of memory called the
7211 \i\e{global offset table}, or \i{GOT}; the GOT is situated at a
7212 constant distance from your library's code, so if you can find out
7213 where your library is loaded (which is typically done using a
7214 \c{CALL} and \c{POP} combination), you can obtain the address of the
7215 GOT, and you can then load the addresses of your variables out of
7216 linker-generated entries in the GOT.
7218 The \e{data} section of a PIC shared library does not have these
7219 restrictions: since the data section is writable, it has to be
7220 copied into memory anyway rather than just paged in from the library
7221 file, so as long as it's being copied it can be relocated too. So
7222 you can put ordinary types of relocation in the data section without
7223 too much worry (but see \k{picglobal} for a caveat).
7226 \S{picgot} Obtaining the Address of the GOT
7228 Each code module in your shared library should define the GOT as an
7231 \c extern _GLOBAL_OFFSET_TABLE_ ; in ELF
7232 \c extern __GLOBAL_OFFSET_TABLE_ ; in BSD a.out
7234 At the beginning of any function in your shared library which plans
7235 to access your data or BSS sections, you must first calculate the
7236 address of the GOT. This is typically done by writing the function
7245 \c add ebx,_GLOBAL_OFFSET_TABLE_+$$-.get_GOT wrt ..gotpc
7247 \c ; the function body comes here
7254 (For BSD, again, the symbol \c{_GLOBAL_OFFSET_TABLE} requires a
7255 second leading underscore.)
7257 The first two lines of this function are simply the standard C
7258 prologue to set up a stack frame, and the last three lines are
7259 standard C function epilogue. The third line, and the fourth to last
7260 line, save and restore the \c{EBX} register, because PIC shared
7261 libraries use this register to store the address of the GOT.
7263 The interesting bit is the \c{CALL} instruction and the following
7264 two lines. The \c{CALL} and \c{POP} combination obtains the address
7265 of the label \c{.get_GOT}, without having to know in advance where
7266 the program was loaded (since the \c{CALL} instruction is encoded
7267 relative to the current position). The \c{ADD} instruction makes use
7268 of one of the special PIC relocation types: \i{GOTPC relocation}.
7269 With the \i\c{WRT ..gotpc} qualifier specified, the symbol
7270 referenced (here \c{_GLOBAL_OFFSET_TABLE_}, the special symbol
7271 assigned to the GOT) is given as an offset from the beginning of the
7272 section. (Actually, \c{ELF} encodes it as the offset from the operand
7273 field of the \c{ADD} instruction, but NASM simplifies this
7274 deliberately, so you do things the same way for both \c{ELF} and
7275 \c{BSD}.) So the instruction then \e{adds} the beginning of the section,
7276 to get the real address of the GOT, and subtracts the value of
7277 \c{.get_GOT} which it knows is in \c{EBX}. Therefore, by the time
7278 that instruction has finished, \c{EBX} contains the address of the GOT.
7280 If you didn't follow that, don't worry: it's never necessary to
7281 obtain the address of the GOT by any other means, so you can put
7282 those three instructions into a macro and safely ignore them:
7289 \c add ebx,_GLOBAL_OFFSET_TABLE_+$$-%%getgot wrt ..gotpc
7293 \S{piclocal} Finding Your Local Data Items
7295 Having got the GOT, you can then use it to obtain the addresses of
7296 your data items. Most variables will reside in the sections you have
7297 declared; they can be accessed using the \I{GOTOFF
7298 relocation}\c{..gotoff} special \I\c{WRT ..gotoff}\c{WRT} type. The
7299 way this works is like this:
7301 \c lea eax,[ebx+myvar wrt ..gotoff]
7303 The expression \c{myvar wrt ..gotoff} is calculated, when the shared
7304 library is linked, to be the offset to the local variable \c{myvar}
7305 from the beginning of the GOT. Therefore, adding it to \c{EBX} as
7306 above will place the real address of \c{myvar} in \c{EAX}.
7308 If you declare variables as \c{GLOBAL} without specifying a size for
7309 them, they are shared between code modules in the library, but do
7310 not get exported from the library to the program that loaded it.
7311 They will still be in your ordinary data and BSS sections, so you
7312 can access them in the same way as local variables, using the above
7313 \c{..gotoff} mechanism.
7315 Note that due to a peculiarity of the way BSD \c{a.out} format
7316 handles this relocation type, there must be at least one non-local
7317 symbol in the same section as the address you're trying to access.
7320 \S{picextern} Finding External and Common Data Items
7322 If your library needs to get at an external variable (external to
7323 the \e{library}, not just to one of the modules within it), you must
7324 use the \I{GOT relocations}\I\c{WRT ..got}\c{..got} type to get at
7325 it. The \c{..got} type, instead of giving you the offset from the
7326 GOT base to the variable, gives you the offset from the GOT base to
7327 a GOT \e{entry} containing the address of the variable. The linker
7328 will set up this GOT entry when it builds the library, and the
7329 dynamic linker will place the correct address in it at load time. So
7330 to obtain the address of an external variable \c{extvar} in \c{EAX},
7333 \c mov eax,[ebx+extvar wrt ..got]
7335 This loads the address of \c{extvar} out of an entry in the GOT. The
7336 linker, when it builds the shared library, collects together every
7337 relocation of type \c{..got}, and builds the GOT so as to ensure it
7338 has every necessary entry present.
7340 Common variables must also be accessed in this way.
7343 \S{picglobal} Exporting Symbols to the Library User
7345 If you want to export symbols to the user of the library, you have
7346 to declare whether they are functions or data, and if they are data,
7347 you have to give the size of the data item. This is because the
7348 dynamic linker has to build \I{PLT}\i{procedure linkage table}
7349 entries for any exported functions, and also moves exported data
7350 items away from the library's data section in which they were
7353 So to export a function to users of the library, you must use
7355 \c global func:function ; declare it as a function
7361 And to export a data item such as an array, you would have to code
7363 \c global array:data array.end-array ; give the size too
7368 Be careful: If you export a variable to the library user, by
7369 declaring it as \c{GLOBAL} and supplying a size, the variable will
7370 end up living in the data section of the main program, rather than
7371 in your library's data section, where you declared it. So you will
7372 have to access your own global variable with the \c{..got} mechanism
7373 rather than \c{..gotoff}, as if it were external (which,
7374 effectively, it has become).
7376 Equally, if you need to store the address of an exported global in
7377 one of your data sections, you can't do it by means of the standard
7380 \c dataptr: dd global_data_item ; WRONG
7382 NASM will interpret this code as an ordinary relocation, in which
7383 \c{global_data_item} is merely an offset from the beginning of the
7384 \c{.data} section (or whatever); so this reference will end up
7385 pointing at your data section instead of at the exported global
7386 which resides elsewhere.
7388 Instead of the above code, then, you must write
7390 \c dataptr: dd global_data_item wrt ..sym
7392 which makes use of the special \c{WRT} type \I\c{WRT ..sym}\c{..sym}
7393 to instruct NASM to search the symbol table for a particular symbol
7394 at that address, rather than just relocating by section base.
7396 Either method will work for functions: referring to one of your
7397 functions by means of
7399 \c funcptr: dd my_function
7401 will give the user the address of the code you wrote, whereas
7403 \c funcptr: dd my_function wrt ..sym
7405 will give the address of the procedure linkage table for the
7406 function, which is where the calling program will \e{believe} the
7407 function lives. Either address is a valid way to call the function.
7410 \S{picproc} Calling Procedures Outside the Library
7412 Calling procedures outside your shared library has to be done by
7413 means of a \i\e{procedure linkage table}, or \i{PLT}. The PLT is
7414 placed at a known offset from where the library is loaded, so the
7415 library code can make calls to the PLT in a position-independent
7416 way. Within the PLT there is code to jump to offsets contained in
7417 the GOT, so function calls to other shared libraries or to routines
7418 in the main program can be transparently passed off to their real
7421 To call an external routine, you must use another special PIC
7422 relocation type, \I{PLT relocations}\i\c{WRT ..plt}. This is much
7423 easier than the GOT-based ones: you simply replace calls such as
7424 \c{CALL printf} with the PLT-relative version \c{CALL printf WRT
7428 \S{link} Generating the Library File
7430 Having written some code modules and assembled them to \c{.o} files,
7431 you then generate your shared library with a command such as
7433 \c ld -shared -o library.so module1.o module2.o # for ELF
7434 \c ld -Bshareable -o library.so module1.o module2.o # for BSD
7436 For ELF, if your shared library is going to reside in system
7437 directories such as \c{/usr/lib} or \c{/lib}, it is usually worth
7438 using the \i\c{-soname} flag to the linker, to store the final
7439 library file name, with a version number, into the library:
7441 \c ld -shared -soname library.so.1 -o library.so.1.2 *.o
7443 You would then copy \c{library.so.1.2} into the library directory,
7444 and create \c{library.so.1} as a symbolic link to it.
7447 \C{mixsize} Mixing 16 and 32 Bit Code
7449 This chapter tries to cover some of the issues, largely related to
7450 unusual forms of addressing and jump instructions, encountered when
7451 writing operating system code such as protected-mode initialisation
7452 routines, which require code that operates in mixed segment sizes,
7453 such as code in a 16-bit segment trying to modify data in a 32-bit
7454 one, or jumps between different-size segments.
7457 \H{mixjump} Mixed-Size Jumps\I{jumps, mixed-size}
7459 \I{operating system, writing}\I{writing operating systems}The most
7460 common form of \i{mixed-size instruction} is the one used when
7461 writing a 32-bit OS: having done your setup in 16-bit mode, such as
7462 loading the kernel, you then have to boot it by switching into
7463 protected mode and jumping to the 32-bit kernel start address. In a
7464 fully 32-bit OS, this tends to be the \e{only} mixed-size
7465 instruction you need, since everything before it can be done in pure
7466 16-bit code, and everything after it can be pure 32-bit.
7468 This jump must specify a 48-bit far address, since the target
7469 segment is a 32-bit one. However, it must be assembled in a 16-bit
7470 segment, so just coding, for example,
7472 \c jmp 0x1234:0x56789ABC ; wrong!
7474 will not work, since the offset part of the address will be
7475 truncated to \c{0x9ABC} and the jump will be an ordinary 16-bit far
7478 The Linux kernel setup code gets round the inability of \c{as86} to
7479 generate the required instruction by coding it manually, using
7480 \c{DB} instructions. NASM can go one better than that, by actually
7481 generating the right instruction itself. Here's how to do it right:
7483 \c jmp dword 0x1234:0x56789ABC ; right
7485 \I\c{JMP DWORD}The \c{DWORD} prefix (strictly speaking, it should
7486 come \e{after} the colon, since it is declaring the \e{offset} field
7487 to be a doubleword; but NASM will accept either form, since both are
7488 unambiguous) forces the offset part to be treated as far, in the
7489 assumption that you are deliberately writing a jump from a 16-bit
7490 segment to a 32-bit one.
7492 You can do the reverse operation, jumping from a 32-bit segment to a
7493 16-bit one, by means of the \c{WORD} prefix:
7495 \c jmp word 0x8765:0x4321 ; 32 to 16 bit
7497 If the \c{WORD} prefix is specified in 16-bit mode, or the \c{DWORD}
7498 prefix in 32-bit mode, they will be ignored, since each is
7499 explicitly forcing NASM into a mode it was in anyway.
7502 \H{mixaddr} Addressing Between Different-Size Segments\I{addressing,
7503 mixed-size}\I{mixed-size addressing}
7505 If your OS is mixed 16 and 32-bit, or if you are writing a DOS
7506 extender, you are likely to have to deal with some 16-bit segments
7507 and some 32-bit ones. At some point, you will probably end up
7508 writing code in a 16-bit segment which has to access data in a
7509 32-bit segment, or vice versa.
7511 If the data you are trying to access in a 32-bit segment lies within
7512 the first 64K of the segment, you may be able to get away with using
7513 an ordinary 16-bit addressing operation for the purpose; but sooner
7514 or later, you will want to do 32-bit addressing from 16-bit mode.
7516 The easiest way to do this is to make sure you use a register for
7517 the address, since any effective address containing a 32-bit
7518 register is forced to be a 32-bit address. So you can do
7520 \c mov eax,offset_into_32_bit_segment_specified_by_fs
7521 \c mov dword [fs:eax],0x11223344
7523 This is fine, but slightly cumbersome (since it wastes an
7524 instruction and a register) if you already know the precise offset
7525 you are aiming at. The x86 architecture does allow 32-bit effective
7526 addresses to specify nothing but a 4-byte offset, so why shouldn't
7527 NASM be able to generate the best instruction for the purpose?
7529 It can. As in \k{mixjump}, you need only prefix the address with the
7530 \c{DWORD} keyword, and it will be forced to be a 32-bit address:
7532 \c mov dword [fs:dword my_offset],0x11223344
7534 Also as in \k{mixjump}, NASM is not fussy about whether the
7535 \c{DWORD} prefix comes before or after the segment override, so
7536 arguably a nicer-looking way to code the above instruction is
7538 \c mov dword [dword fs:my_offset],0x11223344
7540 Don't confuse the \c{DWORD} prefix \e{outside} the square brackets,
7541 which controls the size of the data stored at the address, with the
7542 one \c{inside} the square brackets which controls the length of the
7543 address itself. The two can quite easily be different:
7545 \c mov word [dword 0x12345678],0x9ABC
7547 This moves 16 bits of data to an address specified by a 32-bit
7550 You can also specify \c{WORD} or \c{DWORD} prefixes along with the
7551 \c{FAR} prefix to indirect far jumps or calls. For example:
7553 \c call dword far [fs:word 0x4321]
7555 This instruction contains an address specified by a 16-bit offset;
7556 it loads a 48-bit far pointer from that (16-bit segment and 32-bit
7557 offset), and calls that address.
7560 \H{mixother} Other Mixed-Size Instructions
7562 The other way you might want to access data might be using the
7563 string instructions (\c{LODSx}, \c{STOSx} and so on) or the
7564 \c{XLATB} instruction. These instructions, since they take no
7565 parameters, might seem to have no easy way to make them perform
7566 32-bit addressing when assembled in a 16-bit segment.
7568 This is the purpose of NASM's \i\c{a16}, \i\c{a32} and \i\c{a64} prefixes. If
7569 you are coding \c{LODSB} in a 16-bit segment but it is supposed to
7570 be accessing a string in a 32-bit segment, you should load the
7571 desired address into \c{ESI} and then code
7575 The prefix forces the addressing size to 32 bits, meaning that
7576 \c{LODSB} loads from \c{[DS:ESI]} instead of \c{[DS:SI]}. To access
7577 a string in a 16-bit segment when coding in a 32-bit one, the
7578 corresponding \c{a16} prefix can be used.
7580 The \c{a16}, \c{a32} and \c{a64} prefixes can be applied to any instruction
7581 in NASM's instruction table, but most of them can generate all the
7582 useful forms without them. The prefixes are necessary only for
7583 instructions with implicit addressing:
7584 \# \c{CMPSx} (\k{insCMPSB}),
7585 \# \c{SCASx} (\k{insSCASB}), \c{LODSx} (\k{insLODSB}), \c{STOSx}
7586 \# (\k{insSTOSB}), \c{MOVSx} (\k{insMOVSB}), \c{INSx} (\k{insINSB}),
7587 \# \c{OUTSx} (\k{insOUTSB}), and \c{XLATB} (\k{insXLATB}).
7588 \c{CMPSx}, \c{SCASx}, \c{LODSx}, \c{STOSx}, \c{MOVSx}, \c{INSx},
7589 \c{OUTSx}, and \c{XLATB}.
7591 various push and pop instructions (\c{PUSHA} and \c{POPF} as well as
7592 the more usual \c{PUSH} and \c{POP}) can accept \c{a16}, \c{a32} or \c{a64}
7593 prefixes to force a particular one of \c{SP}, \c{ESP} or \c{RSP} to be used
7594 as a stack pointer, in case the stack segment in use is a different
7595 size from the code segment.
7597 \c{PUSH} and \c{POP}, when applied to segment registers in 32-bit
7598 mode, also have the slightly odd behaviour that they push and pop 4
7599 bytes at a time, of which the top two are ignored and the bottom two
7600 give the value of the segment register being manipulated. To force
7601 the 16-bit behaviour of segment-register push and pop instructions,
7602 you can use the operand-size prefix \i\c{o16}:
7607 This code saves a doubleword of stack space by fitting two segment
7608 registers into the space which would normally be consumed by pushing
7611 (You can also use the \i\c{o32} prefix to force the 32-bit behaviour
7612 when in 16-bit mode, but this seems less useful.)
7615 \C{64bit} Writing 64-bit Code (Unix, Win64)
7617 This chapter attempts to cover some of the common issues involved when
7618 writing 64-bit code, to run under \i{Win64} or Unix. It covers how to
7619 write assembly code to interface with 64-bit C routines, and how to
7620 write position-independent code for shared libraries.
7622 All 64-bit code uses a flat memory model, since segmentation is not
7623 available in 64-bit mode. The one exception is the \c{FS} and \c{GS}
7624 registers, which still add their bases.
7626 Position independence in 64-bit mode is significantly simpler, since
7627 the processor supports \c{RIP}-relative addressing directly; see the
7628 \c{REL} keyword (\k{effaddr}). On most 64-bit platforms, it is
7629 probably desirable to make that the default, using the directive
7630 \c{DEFAULT REL} (\k{default}).
7632 64-bit programming is relatively similar to 32-bit programming, but
7633 of course pointers are 64 bits long; additionally, all existing
7634 platforms pass arguments in registers rather than on the stack.
7635 Furthermore, 64-bit platforms use SSE2 by default for floating point.
7636 Please see the ABI documentation for your platform.
7638 64-bit platforms differ in the sizes of the fundamental datatypes, not
7639 just from 32-bit platforms but from each other. If a specific size
7640 data type is desired, it is probably best to use the types defined in
7641 the Standard C header \c{<inttypes.h>}.
7643 In 64-bit mode, the default instruction size is still 32 bits. When
7644 loading a value into a 32-bit register (but not an 8- or 16-bit
7645 register), the upper 32 bits of the corresponding 64-bit register are
7648 \H{reg64} Register Names in 64-bit Mode
7650 NASM uses the following names for general-purpose registers in 64-bit
7651 mode, for 8-, 16-, 32- and 64-bit references, respecitively:
7653 \c AL/AH, CL/CH, DL/DH, BL/BH, SPL, BPL, SIL, DIL, R8B-R15B
7654 \c AX, CX, DX, BX, SP, BP, SI, DI, R8W-R15W
7655 \c EAX, ECX, EDX, EBX, ESP, EBP, ESI, EDI, R8D-R15D
7656 \c RAX, RCX, RDX, RBX, RSP, RBP, RSI, RDI, R8-R15
7658 This is consistent with the AMD documentation and most other
7659 assemblers. The Intel documentation, however, uses the names
7660 \c{R8L-R15L} for 8-bit references to the higher registers. It is
7661 possible to use those names by definiting them as macros; similarly,
7662 if one wants to use numeric names for the low 8 registers, define them
7663 as macros. The standard macro package \c{altreg} (see \k{pkg_altreg})
7664 can be used for this purpose.
7666 \H{id64} Immediates and Displacements in 64-bit Mode
7668 In 64-bit mode, immediates and displacements are generally only 32
7669 bits wide. NASM will therefore truncate most displacements and
7670 immediates to 32 bits.
7672 The only instruction which takes a full \i{64-bit immediate} is:
7676 NASM will produce this instruction whenever the programmer uses
7677 \c{MOV} with an immediate into a 64-bit register. If this is not
7678 desirable, simply specify the equivalent 32-bit register, which will
7679 be automatically zero-extended by the processor, or specify the
7680 immediate as \c{DWORD}:
7682 \c mov rax,foo ; 64-bit immediate
7683 \c mov rax,qword foo ; (identical)
7684 \c mov eax,foo ; 32-bit immediate, zero-extended
7685 \c mov rax,dword foo ; 32-bit immediate, sign-extended
7687 The length of these instructions are 10, 5 and 7 bytes, respectively.
7689 The only instructions which take a full \I{64-bit displacement}64-bit
7690 \e{displacement} is loading or storing, using \c{MOV}, \c{AL}, \c{AX},
7691 \c{EAX} or \c{RAX} (but no other registers) to an absolute 64-bit address.
7692 Since this is a relatively rarely used instruction (64-bit code generally uses
7693 relative addressing), the programmer has to explicitly declare the
7694 displacement size as \c{QWORD}:
7698 \c mov eax,[foo] ; 32-bit absolute disp, sign-extended
7699 \c mov eax,[a32 foo] ; 32-bit absolute disp, zero-extended
7700 \c mov eax,[qword foo] ; 64-bit absolute disp
7704 \c mov eax,[foo] ; 32-bit relative disp
7705 \c mov eax,[a32 foo] ; d:o, address truncated to 32 bits(!)
7706 \c mov eax,[qword foo] ; error
7707 \c mov eax,[abs qword foo] ; 64-bit absolute disp
7709 A sign-extended absolute displacement can access from -2 GB to +2 GB;
7710 a zero-extended absolute displacement can access from 0 to 4 GB.
7712 \H{unix64} Interfacing to 64-bit C Programs (Unix)
7714 On Unix, the 64-bit ABI is defined by the document:
7716 \W{http://www.nasm.us/links/unix64abi}\c{http://www.nasm.us/links/unix64abi}
7718 Although written for AT&T-syntax assembly, the concepts apply equally
7719 well for NASM-style assembly. What follows is a simplified summary.
7721 The first six integer arguments (from the left) are passed in \c{RDI},
7722 \c{RSI}, \c{RDX}, \c{RCX}, \c{R8}, and \c{R9}, in that order.
7723 Additional integer arguments are passed on the stack. These
7724 registers, plus \c{RAX}, \c{R10} and \c{R11} are destroyed by function
7725 calls, and thus are available for use by the function without saving.
7727 Integer return values are passed in \c{RAX} and \c{RDX}, in that order.
7729 Floating point is done using SSE registers, except for \c{long
7730 double}. Floating-point arguments are passed in \c{XMM0} to \c{XMM7};
7731 return is \c{XMM0} and \c{XMM1}. \c{long double} are passed on the
7732 stack, and returned in \c{ST0} and \c{ST1}.
7734 All SSE and x87 registers are destroyed by function calls.
7736 On 64-bit Unix, \c{long} is 64 bits.
7738 Integer and SSE register arguments are counted separately, so for the case of
7740 \c void foo(long a, double b, int c)
7742 \c{a} is passed in \c{RDI}, \c{b} in \c{XMM0}, and \c{c} in \c{ESI}.
7744 \H{win64} Interfacing to 64-bit C Programs (Win64)
7746 The Win64 ABI is described at:
7748 \W{http://www.nasm.us/links/win64abi}\c{http://www.nasm.us/links/win64abi}
7750 What follows is a simplified summary.
7752 The first four integer arguments are passed in \c{RCX}, \c{RDX},
7753 \c{R8} and \c{R9}, in that order. Additional integer arguments are
7754 passed on the stack. These registers, plus \c{RAX}, \c{R10} and
7755 \c{R11} are destroyed by function calls, and thus are available for
7756 use by the function without saving.
7758 Integer return values are passed in \c{RAX} only.
7760 Floating point is done using SSE registers, except for \c{long
7761 double}. Floating-point arguments are passed in \c{XMM0} to \c{XMM3};
7762 return is \c{XMM0} only.
7764 On Win64, \c{long} is 32 bits; \c{long long} or \c{_int64} is 64 bits.
7766 Integer and SSE register arguments are counted together, so for the case of
7768 \c void foo(long long a, double b, int c)
7770 \c{a} is passed in \c{RCX}, \c{b} in \c{XMM1}, and \c{c} in \c{R8D}.
7772 \C{trouble} Troubleshooting
7774 This chapter describes some of the common problems that users have
7775 been known to encounter with NASM, and answers them. It also gives
7776 instructions for reporting bugs in NASM if you find a difficulty
7777 that isn't listed here.
7780 \H{problems} Common Problems
7782 \S{inefficient} NASM Generates \i{Inefficient Code}
7784 We sometimes get `bug' reports about NASM generating inefficient, or
7785 even `wrong', code on instructions such as \c{ADD ESP,8}. This is a
7786 deliberate design feature, connected to predictability of output:
7787 NASM, on seeing \c{ADD ESP,8}, will generate the form of the
7788 instruction which leaves room for a 32-bit offset. You need to code
7789 \I\c{BYTE}\c{ADD ESP,BYTE 8} if you want the space-efficient form of
7790 the instruction. This isn't a bug, it's user error: if you prefer to
7791 have NASM produce the more efficient code automatically enable
7792 optimization with the \c{-O} option (see \k{opt-O}).
7795 \S{jmprange} My Jumps are Out of Range\I{out of range, jumps}
7797 Similarly, people complain that when they issue \i{conditional
7798 jumps} (which are \c{SHORT} by default) that try to jump too far,
7799 NASM reports `short jump out of range' instead of making the jumps
7802 This, again, is partly a predictability issue, but in fact has a
7803 more practical reason as well. NASM has no means of being told what
7804 type of processor the code it is generating will be run on; so it
7805 cannot decide for itself that it should generate \i\c{Jcc NEAR} type
7806 instructions, because it doesn't know that it's working for a 386 or
7807 above. Alternatively, it could replace the out-of-range short
7808 \c{JNE} instruction with a very short \c{JE} instruction that jumps
7809 over a \c{JMP NEAR}; this is a sensible solution for processors
7810 below a 386, but hardly efficient on processors which have good
7811 branch prediction \e{and} could have used \c{JNE NEAR} instead. So,
7812 once again, it's up to the user, not the assembler, to decide what
7813 instructions should be generated. See \k{opt-O}.
7816 \S{proborg} \i\c{ORG} Doesn't Work
7818 People writing \i{boot sector} programs in the \c{bin} format often
7819 complain that \c{ORG} doesn't work the way they'd like: in order to
7820 place the \c{0xAA55} signature word at the end of a 512-byte boot
7821 sector, people who are used to MASM tend to code
7825 \c ; some boot sector code
7830 This is not the intended use of the \c{ORG} directive in NASM, and
7831 will not work. The correct way to solve this problem in NASM is to
7832 use the \i\c{TIMES} directive, like this:
7836 \c ; some boot sector code
7838 \c TIMES 510-($-$$) DB 0
7841 The \c{TIMES} directive will insert exactly enough zero bytes into
7842 the output to move the assembly point up to 510. This method also
7843 has the advantage that if you accidentally fill your boot sector too
7844 full, NASM will catch the problem at assembly time and report it, so
7845 you won't end up with a boot sector that you have to disassemble to
7846 find out what's wrong with it.
7849 \S{probtimes} \i\c{TIMES} Doesn't Work
7851 The other common problem with the above code is people who write the
7856 by reasoning that \c{$} should be a pure number, just like 510, so
7857 the difference between them is also a pure number and can happily be
7860 NASM is a \e{modular} assembler: the various component parts are
7861 designed to be easily separable for re-use, so they don't exchange
7862 information unnecessarily. In consequence, the \c{bin} output
7863 format, even though it has been told by the \c{ORG} directive that
7864 the \c{.text} section should start at 0, does not pass that
7865 information back to the expression evaluator. So from the
7866 evaluator's point of view, \c{$} isn't a pure number: it's an offset
7867 from a section base. Therefore the difference between \c{$} and 510
7868 is also not a pure number, but involves a section base. Values
7869 involving section bases cannot be passed as arguments to \c{TIMES}.
7871 The solution, as in the previous section, is to code the \c{TIMES}
7874 \c TIMES 510-($-$$) DB 0
7876 in which \c{$} and \c{$$} are offsets from the same section base,
7877 and so their difference is a pure number. This will solve the
7878 problem and generate sensible code.
7881 \H{bugs} \i{Bugs}\I{reporting bugs}
7883 We have never yet released a version of NASM with any \e{known}
7884 bugs. That doesn't usually stop there being plenty we didn't know
7885 about, though. Any that you find should be reported firstly via the
7887 \W{http://www.nasm.us/}\c{http://www.nasm.us/}
7888 (click on "Bug Tracker"), or if that fails then through one of the
7889 contacts in \k{contact}.
7891 Please read \k{qstart} first, and don't report the bug if it's
7892 listed in there as a deliberate feature. (If you think the feature
7893 is badly thought out, feel free to send us reasons why you think it
7894 should be changed, but don't just send us mail saying `This is a
7895 bug' if the documentation says we did it on purpose.) Then read
7896 \k{problems}, and don't bother reporting the bug if it's listed
7899 If you do report a bug, \e{please} give us all of the following
7902 \b What operating system you're running NASM under. DOS, Linux,
7903 NetBSD, Win16, Win32, VMS (I'd be impressed), whatever.
7905 \b If you're running NASM under DOS or Win32, tell us whether you've
7906 compiled your own executable from the DOS source archive, or whether
7907 you were using the standard distribution binaries out of the
7908 archive. If you were using a locally built executable, try to
7909 reproduce the problem using one of the standard binaries, as this
7910 will make it easier for us to reproduce your problem prior to fixing
7913 \b Which version of NASM you're using, and exactly how you invoked
7914 it. Give us the precise command line, and the contents of the
7915 \c{NASMENV} environment variable if any.
7917 \b Which versions of any supplementary programs you're using, and
7918 how you invoked them. If the problem only becomes visible at link
7919 time, tell us what linker you're using, what version of it you've
7920 got, and the exact linker command line. If the problem involves
7921 linking against object files generated by a compiler, tell us what
7922 compiler, what version, and what command line or options you used.
7923 (If you're compiling in an IDE, please try to reproduce the problem
7924 with the command-line version of the compiler.)
7926 \b If at all possible, send us a NASM source file which exhibits the
7927 problem. If this causes copyright problems (e.g. you can only
7928 reproduce the bug in restricted-distribution code) then bear in mind
7929 the following two points: firstly, we guarantee that any source code
7930 sent to us for the purposes of debugging NASM will be used \e{only}
7931 for the purposes of debugging NASM, and that we will delete all our
7932 copies of it as soon as we have found and fixed the bug or bugs in
7933 question; and secondly, we would prefer \e{not} to be mailed large
7934 chunks of code anyway. The smaller the file, the better. A
7935 three-line sample file that does nothing useful \e{except}
7936 demonstrate the problem is much easier to work with than a
7937 fully fledged ten-thousand-line program. (Of course, some errors
7938 \e{do} only crop up in large files, so this may not be possible.)
7940 \b A description of what the problem actually \e{is}. `It doesn't
7941 work' is \e{not} a helpful description! Please describe exactly what
7942 is happening that shouldn't be, or what isn't happening that should.
7943 Examples might be: `NASM generates an error message saying Line 3
7944 for an error that's actually on Line 5'; `NASM generates an error
7945 message that I believe it shouldn't be generating at all'; `NASM
7946 fails to generate an error message that I believe it \e{should} be
7947 generating'; `the object file produced from this source code crashes
7948 my linker'; `the ninth byte of the output file is 66 and I think it
7949 should be 77 instead'.
7951 \b If you believe the output file from NASM to be faulty, send it to
7952 us. That allows us to determine whether our own copy of NASM
7953 generates the same file, or whether the problem is related to
7954 portability issues between our development platforms and yours. We
7955 can handle binary files mailed to us as MIME attachments, uuencoded,
7956 and even BinHex. Alternatively, we may be able to provide an FTP
7957 site you can upload the suspect files to; but mailing them is easier
7960 \b Any other information or data files that might be helpful. If,
7961 for example, the problem involves NASM failing to generate an object
7962 file while TASM can generate an equivalent file without trouble,
7963 then send us \e{both} object files, so we can see what TASM is doing
7964 differently from us.
7967 \A{ndisasm} \i{Ndisasm}
7969 The Netwide Disassembler, NDISASM
7971 \H{ndisintro} Introduction
7974 The Netwide Disassembler is a small companion program to the Netwide
7975 Assembler, NASM. It seemed a shame to have an x86 assembler,
7976 complete with a full instruction table, and not make as much use of
7977 it as possible, so here's a disassembler which shares the
7978 instruction table (and some other bits of code) with NASM.
7980 The Netwide Disassembler does nothing except to produce
7981 disassemblies of \e{binary} source files. NDISASM does not have any
7982 understanding of object file formats, like \c{objdump}, and it will
7983 not understand \c{DOS .EXE} files like \c{debug} will. It just
7987 \H{ndisstart} Getting Started: Installation
7989 See \k{install} for installation instructions. NDISASM, like NASM,
7990 has a \c{man page} which you may want to put somewhere useful, if you
7991 are on a Unix system.
7994 \H{ndisrun} Running NDISASM
7996 To disassemble a file, you will typically use a command of the form
7998 \c ndisasm -b {16|32|64} filename
8000 NDISASM can disassemble 16-, 32- or 64-bit code equally easily,
8001 provided of course that you remember to specify which it is to work
8002 with. If no \i\c{-b} switch is present, NDISASM works in 16-bit mode
8003 by default. The \i\c{-u} switch (for USE32) also invokes 32-bit mode.
8005 Two more command line options are \i\c{-r} which reports the version
8006 number of NDISASM you are running, and \i\c{-h} which gives a short
8007 summary of command line options.
8010 \S{ndiscom} COM Files: Specifying an Origin
8012 To disassemble a \c{DOS .COM} file correctly, a disassembler must assume
8013 that the first instruction in the file is loaded at address \c{0x100},
8014 rather than at zero. NDISASM, which assumes by default that any file
8015 you give it is loaded at zero, will therefore need to be informed of
8018 The \i\c{-o} option allows you to declare a different origin for the
8019 file you are disassembling. Its argument may be expressed in any of
8020 the NASM numeric formats: decimal by default, if it begins with `\c{$}'
8021 or `\c{0x}' or ends in `\c{H}' it's \c{hex}, if it ends in `\c{Q}' it's
8022 \c{octal}, and if it ends in `\c{B}' it's \c{binary}.
8024 Hence, to disassemble a \c{.COM} file:
8026 \c ndisasm -o100h filename.com
8031 \S{ndissync} Code Following Data: Synchronisation
8033 Suppose you are disassembling a file which contains some data which
8034 isn't machine code, and \e{then} contains some machine code. NDISASM
8035 will faithfully plough through the data section, producing machine
8036 instructions wherever it can (although most of them will look
8037 bizarre, and some may have unusual prefixes, e.g. `\c{FS OR AX,0x240A}'),
8038 and generating `DB' instructions ever so often if it's totally stumped.
8039 Then it will reach the code section.
8041 Supposing NDISASM has just finished generating a strange machine
8042 instruction from part of the data section, and its file position is
8043 now one byte \e{before} the beginning of the code section. It's
8044 entirely possible that another spurious instruction will get
8045 generated, starting with the final byte of the data section, and
8046 then the correct first instruction in the code section will not be
8047 seen because the starting point skipped over it. This isn't really
8050 To avoid this, you can specify a `\i\c{synchronisation}' point, or indeed
8051 as many synchronisation points as you like (although NDISASM can
8052 only handle 2147483647 sync points internally). The definition of a sync
8053 point is this: NDISASM guarantees to hit sync points exactly during
8054 disassembly. If it is thinking about generating an instruction which
8055 would cause it to jump over a sync point, it will discard that
8056 instruction and output a `\c{db}' instead. So it \e{will} start
8057 disassembly exactly from the sync point, and so you \e{will} see all
8058 the instructions in your code section.
8060 Sync points are specified using the \i\c{-s} option: they are measured
8061 in terms of the program origin, not the file position. So if you
8062 want to synchronize after 32 bytes of a \c{.COM} file, you would have to
8065 \c ndisasm -o100h -s120h file.com
8069 \c ndisasm -o100h -s20h file.com
8071 As stated above, you can specify multiple sync markers if you need
8072 to, just by repeating the \c{-s} option.
8075 \S{ndisisync} Mixed Code and Data: Automatic (Intelligent) Synchronisation
8078 Suppose you are disassembling the boot sector of a \c{DOS} floppy (maybe
8079 it has a virus, and you need to understand the virus so that you
8080 know what kinds of damage it might have done you). Typically, this
8081 will contain a \c{JMP} instruction, then some data, then the rest of the
8082 code. So there is a very good chance of NDISASM being \e{misaligned}
8083 when the data ends and the code begins. Hence a sync point is
8086 On the other hand, why should you have to specify the sync point
8087 manually? What you'd do in order to find where the sync point would
8088 be, surely, would be to read the \c{JMP} instruction, and then to use
8089 its target address as a sync point. So can NDISASM do that for you?
8091 The answer, of course, is yes: using either of the synonymous
8092 switches \i\c{-a} (for automatic sync) or \i\c{-i} (for intelligent
8093 sync) will enable \c{auto-sync} mode. Auto-sync mode automatically
8094 generates a sync point for any forward-referring PC-relative jump or
8095 call instruction that NDISASM encounters. (Since NDISASM is one-pass,
8096 if it encounters a PC-relative jump whose target has already been
8097 processed, there isn't much it can do about it...)
8099 Only PC-relative jumps are processed, since an absolute jump is
8100 either through a register (in which case NDISASM doesn't know what
8101 the register contains) or involves a segment address (in which case
8102 the target code isn't in the same segment that NDISASM is working
8103 in, and so the sync point can't be placed anywhere useful).
8105 For some kinds of file, this mechanism will automatically put sync
8106 points in all the right places, and save you from having to place
8107 any sync points manually. However, it should be stressed that
8108 auto-sync mode is \e{not} guaranteed to catch all the sync points, and
8109 you may still have to place some manually.
8111 Auto-sync mode doesn't prevent you from declaring manual sync
8112 points: it just adds automatically generated ones to the ones you
8113 provide. It's perfectly feasible to specify \c{-i} \e{and} some \c{-s}
8116 Another caveat with auto-sync mode is that if, by some unpleasant
8117 fluke, something in your data section should disassemble to a
8118 PC-relative call or jump instruction, NDISASM may obediently place a
8119 sync point in a totally random place, for example in the middle of
8120 one of the instructions in your code section. So you may end up with
8121 a wrong disassembly even if you use auto-sync. Again, there isn't
8122 much I can do about this. If you have problems, you'll have to use
8123 manual sync points, or use the \c{-k} option (documented below) to
8124 suppress disassembly of the data area.
8127 \S{ndisother} Other Options
8129 The \i\c{-e} option skips a header on the file, by ignoring the first N
8130 bytes. This means that the header is \e{not} counted towards the
8131 disassembly offset: if you give \c{-e10 -o10}, disassembly will start
8132 at byte 10 in the file, and this will be given offset 10, not 20.
8134 The \i\c{-k} option is provided with two comma-separated numeric
8135 arguments, the first of which is an assembly offset and the second
8136 is a number of bytes to skip. This \e{will} count the skipped bytes
8137 towards the assembly offset: its use is to suppress disassembly of a
8138 data section which wouldn't contain anything you wanted to see
8142 \H{ndisbugs} Bugs and Improvements
8144 There are no known bugs. However, any you find, with patches if
8145 possible, should be sent to
8146 \W{mailto:nasm-bugs@lists.sourceforge.net}\c{nasm-bugs@lists.sourceforge.net}, or to the
8148 \W{http://www.nasm.us/}\c{http://www.nasm.us/}
8149 and we'll try to fix them. Feel free to send contributions and
8150 new features as well.
8152 \A{inslist} \i{Instruction List}
8154 \H{inslistintro} Introduction
8156 The following sections show the instructions which NASM currently supports. For each
8157 instruction, there is a separate entry for each supported addressing mode. The third
8158 column shows the processor type in which the instruction was introduced and,
8159 when appropriate, one or more usage flags.
8163 \A{changelog} \i{NASM Version History}