2 \# Source code to NASM documentation
4 \M{category}{Programming}
5 \M{title}{NASM - The Netwide Assembler}
7 \M{author}{The NASM Development Team}
8 \M{license}{All rights reserved. This document is redistributable under the license given in the file "COPYING" distributed in the NASM archive.}
9 \M{summary}{This file documents NASM, the Netwide Assembler: an assembler targetting the Intel x86 series of processors, with portable source.}
12 \M{infotitle}{The Netwide Assembler for x86}
13 \M{epslogo}{nasmlogo.eps}
19 \IR{-MG} \c{-MG} option
20 \IR{-On} \c{-On} option
39 \IR{!=} \c{!=} operator
40 \IR{$, here} \c{$}, Here token
41 \IR{$, prefix} \c{$}, prefix
44 \IR{%%} \c{%%} operator
45 \IR{%+1} \c{%+1} and \c{%-1} syntax
47 \IR{%0} \c{%0} parameter count
49 \IR{&&} \c{&&} operator
51 \IR{..@} \c{..@} symbol prefix
53 \IR{//} \c{//} operator
55 \IR{<<} \c{<<} operator
56 \IR{<=} \c{<=} operator
57 \IR{<>} \c{<>} operator
59 \IR{==} \c{==} operator
61 \IR{>=} \c{>=} operator
62 \IR{>>} \c{>>} operator
63 \IR{?} \c{?} MASM syntax
65 \IR{^^} \c{^^} operator
67 \IR{||} \c{||} operator
69 \IR{%$} \c{%$} and \c{%$$} prefixes
71 \IR{+ opaddition} \c{+} operator, binary
72 \IR{+ opunary} \c{+} operator, unary
73 \IR{+ modifier} \c{+} modifier
74 \IR{- opsubtraction} \c{-} operator, binary
75 \IR{- opunary} \c{-} operator, unary
76 \IR{! opunary} \c{!} operator, unary
77 \IR{alignment, in bin sections} alignment, in \c{bin} sections
78 \IR{alignment, in elf sections} alignment, in \c{elf} sections
79 \IR{alignment, in win32 sections} alignment, in \c{win32} sections
80 \IR{alignment, of elf common variables} alignment, of \c{elf} common
82 \IR{alignment, in obj sections} alignment, in \c{obj} sections
83 \IR{a.out, bsd version} \c{a.out}, BSD version
84 \IR{a.out, linux version} \c{a.out}, Linux version
85 \IR{autoconf} Autoconf
87 \IR{bitwise and} bitwise AND
88 \IR{bitwise or} bitwise OR
89 \IR{bitwise xor} bitwise XOR
90 \IR{block ifs} block IFs
91 \IR{borland pascal} Borland, Pascal
92 \IR{borland's win32 compilers} Borland, Win32 compilers
93 \IR{braces, after % sign} braces, after \c{%} sign
95 \IR{c calling convention} C calling convention
96 \IR{c symbol names} C symbol names
97 \IA{critical expressions}{critical expression}
98 \IA{command line}{command-line}
99 \IA{case sensitivity}{case sensitive}
100 \IA{case-sensitive}{case sensitive}
101 \IA{case-insensitive}{case sensitive}
102 \IA{character constants}{character constant}
103 \IR{common object file format} Common Object File Format
104 \IR{common variables, alignment in elf} common variables, alignment
106 \IR{common, elf extensions to} \c{COMMON}, \c{elf} extensions to
107 \IR{common, obj extensions to} \c{COMMON}, \c{obj} extensions to
108 \IR{declaring structure} declaring structures
109 \IR{default-wrt mechanism} default-\c{WRT} mechanism
112 \IR{dll symbols, exporting} DLL symbols, exporting
113 \IR{dll symbols, importing} DLL symbols, importing
115 \IR{dos archive} DOS archive
116 \IR{dos source archive} DOS source archive
117 \IA{effective address}{effective addresses}
118 \IA{effective-address}{effective addresses}
120 \IR{elf, 16-bit code and} ELF, 16-bit code and
121 \IR{elf shared libraries} ELF, shared libraries
122 \IR{executable and linkable format} Executable and Linkable Format
123 \IR{extern, obj extensions to} \c{EXTERN}, \c{obj} extensions to
124 \IR{extern, rdf extensions to} \c{EXTERN}, \c{rdf} extensions to
126 \IR{freelink} FreeLink
127 \IR{functions, c calling convention} functions, C calling convention
128 \IR{functions, pascal calling convention} functions, Pascal calling
130 \IR{global, aoutb extensions to} \c{GLOBAL}, \c{aoutb} extensions to
131 \IR{global, elf extensions to} \c{GLOBAL}, \c{elf} extensions to
132 \IR{global, rdf extensions to} \c{GLOBAL}, \c{rdf} extensions to
134 \IR{got relocations} \c{GOT} relocations
135 \IR{gotoff relocation} \c{GOTOFF} relocations
136 \IR{gotpc relocation} \c{GOTPC} relocations
137 \IR{intel number formats} Intel number formats
138 \IR{linux, elf} Linux, ELF
139 \IR{linux, a.out} Linux, \c{a.out}
140 \IR{linux, as86} Linux, \c{as86}
141 \IR{logical and} logical AND
142 \IR{logical or} logical OR
143 \IR{logical xor} logical XOR
145 \IA{memory reference}{memory references}
147 \IA{misc directory}{misc subdirectory}
148 \IR{misc subdirectory} \c{misc} subdirectory
149 \IR{microsoft omf} Microsoft OMF
150 \IR{mmx registers} MMX registers
151 \IA{modr/m}{modr/m byte}
152 \IR{modr/m byte} ModR/M byte
154 \IR{ms-dos device drivers} MS-DOS device drivers
155 \IR{multipush} \c{multipush} macro
157 \IR{nasm version} NASM version
161 \IR{operating system} operating system
163 \IR{pascal calling convention}Pascal calling convention
164 \IR{passes} passes, assembly
169 \IR{plt} \c{PLT} relocations
170 \IA{pre-defining macros}{pre-define}
171 \IA{preprocessor expressions}{preprocessor, expressions}
172 \IA{preprocessor loops}{preprocessor, loops}
173 \IA{preprocessor variables}{preprocessor, variables}
174 \IA{rdoff subdirectory}{rdoff}
175 \IR{rdoff} \c{rdoff} subdirectory
176 \IR{relocatable dynamic object file format} Relocatable Dynamic
178 \IR{relocations, pic-specific} relocations, PIC-specific
179 \IA{repeating}{repeating code}
180 \IR{section alignment, in elf} section alignment, in \c{elf}
181 \IR{section alignment, in bin} section alignment, in \c{bin}
182 \IR{section alignment, in obj} section alignment, in \c{obj}
183 \IR{section alignment, in win32} section alignment, in \c{win32}
184 \IR{section, elf extensions to} \c{SECTION}, \c{elf} extensions to
185 \IR{section, win32 extensions to} \c{SECTION}, \c{win32} extensions to
186 \IR{segment alignment, in bin} segment alignment, in \c{bin}
187 \IR{segment alignment, in obj} segment alignment, in \c{obj}
188 \IR{segment, obj extensions to} \c{SEGMENT}, \c{elf} extensions to
189 \IR{segment names, borland pascal} segment names, Borland Pascal
190 \IR{shift command} \c{shift} command
192 \IR{sib byte} SIB byte
193 \IR{solaris x86} Solaris x86
194 \IA{standard section names}{standardized section names}
195 \IR{symbols, exporting from dlls} symbols, exporting from DLLs
196 \IR{symbols, importing from dlls} symbols, importing from DLLs
197 \IR{test subdirectory} \c{test} subdirectory
199 \IR{underscore, in c symbols} underscore, in C symbols
201 \IA{sco unix}{unix, sco}
202 \IR{unix, sco} Unix, SCO
203 \IA{unix source archive}{unix, source archive}
204 \IR{unix, source archive} Unix, source archive
205 \IA{unix system v}{unix, system v}
206 \IR{unix, system v} Unix, System V
207 \IR{unixware} UnixWare
209 \IR{version number of nasm} version number of NASM
210 \IR{visual c++} Visual C++
211 \IR{www page} WWW page
215 \IR{windows 95} Windows 95
216 \IR{windows nt} Windows NT
217 \# \IC{program entry point}{entry point, program}
218 \# \IC{program entry point}{start point, program}
219 \# \IC{MS-DOS device drivers}{device drivers, MS-DOS}
220 \# \IC{16-bit mode, versus 32-bit mode}{32-bit mode, versus 16-bit mode}
221 \# \IC{c symbol names}{symbol names, in C}
224 \C{intro} Introduction
226 \H{whatsnasm} What Is NASM?
228 The Netwide Assembler, NASM, is an 80x86 and x86-64 assembler designed for
229 portability and modularity. It supports a range of object file
230 formats, including Linux and \c{*BSD} \c{a.out}, \c{ELF}, \c{COFF}, \c{Mach-O},
231 Microsoft 16-bit \c{OBJ}, \c{Win32} and \c{Win64}. It will also output plain
232 binary files. Its syntax is designed to be simple and easy to understand, similar
233 to Intel's but less complex. It supports from the upto and including \c{Pentium},
234 \c{P6}, \c{MMX}, \c{3DNow!}, \c{SSE}, \c{SSE2}, \c{SSE3} and \c{x64} opcodes. NASM has
235 a strong support for macro conventions.
238 \S{yaasm} Why Yet Another Assembler?
240 The Netwide Assembler grew out of an idea on \i\c{comp.lang.asm.x86}
241 (or possibly \i\c{alt.lang.asm} - I forget which), which was
242 essentially that there didn't seem to be a good \e{free} x86-series
243 assembler around, and that maybe someone ought to write one.
245 \b \i\c{a86} is good, but not free, and in particular you don't get any
246 32-bit capability until you pay. It's DOS only, too.
248 \b \i\c{gas} is free, and ports over to DOS and Unix, but it's not
249 very good, since it's designed to be a back end to \i\c{gcc}, which
250 always feeds it correct code. So its error checking is minimal. Also,
251 its syntax is horrible, from the point of view of anyone trying to
252 actually \e{write} anything in it. Plus you can't write 16-bit code in
255 \b \i\c{as86} is specific to Minix and Linux, and (my version at least)
256 doesn't seem to have much (or any) documentation.
258 \b \i\c{MASM} isn't very good, and it's (was) expensive, and it runs only under
261 \b \i\c{TASM} is better, but still strives for MASM compatibility,
262 which means millions of directives and tons of red tape. And its syntax
263 is essentially MASM's, with the contradictions and quirks that
264 entails (although it sorts out some of those by means of Ideal mode.)
265 It's expensive too. And it's DOS-only.
267 So here, for your coding pleasure, is NASM. At present it's
268 still in prototype stage - we don't promise that it can outperform
269 any of these assemblers. But please, \e{please} send us bug reports,
270 fixes, helpful information, and anything else you can get your hands
271 on (and thanks to the many people who've done this already! You all
272 know who you are), and we'll improve it out of all recognition.
276 \S{legal} License Conditions
278 Please see the file \c{COPYING}, supplied as part of any NASM
279 distribution archive, for the \i{license} conditions under which you
280 may use NASM. NASM is now under the so-called GNU Lesser General
281 Public License, LGPL.
284 \H{contact} Contact Information
286 The current version of NASM (since about 0.98.08) is maintained by a
287 team of developers, accessible through the \c{nasm-devel} mailing list
288 (see below for the link).
289 If you want to report a bug, please read \k{bugs} first.
291 NASM has a \i{WWW page} at
292 \W{http://nasm.sourceforge.net}\c{http://nasm.sourceforge.net}. If it's
293 not there, google for us!
296 The original authors are \i{e\-mail}able as
297 \W{mailto:jules@dsf.org.uk}\c{jules@dsf.org.uk} and
298 \W{mailto:anakin@pobox.com}\c{anakin@pobox.com}.
299 The latter is no longer involved in the development team.
301 \i{New releases} of NASM are uploaded to the official sites
302 \W{http://nasm.sourceforge.net}\c{http://nasm.sourceforge.net}
304 \W{ftp://ftp.kernel.org/pub/software/devel/nasm/}\i\c{ftp.kernel.org}
306 \W{ftp://ibiblio.org/pub/Linux/devel/lang/assemblers/}\i\c{ibiblio.org}.
308 Announcements are posted to
309 \W{news:comp.lang.asm.x86}\i\c{comp.lang.asm.x86},
310 \W{news:alt.lang.asm}\i\c{alt.lang.asm} and
311 \W{news:comp.os.linux.announce}\i\c{comp.os.linux.announce}
313 If you want information about NASM beta releases, and the current
314 development status, please subscribe to the \i\c{nasm-devel} email list
316 \W{http://sourceforge.net/projects/nasm}\c{http://sourceforge.net/projects/nasm}.
319 \H{install} Installation
321 \S{instdos} \i{Installing} NASM under MS-\i{DOS} or Windows
323 Once you've obtained the \i{DOS archive} for NASM, \i\c{nasmXXX.zip}
324 (where \c{XXX} denotes the version number of NASM contained in the
325 archive), unpack it into its own directory (for example \c{c:\\nasm}).
327 The archive will contain four executable files: the NASM executable
328 files \i\c{nasm.exe} and \i\c{nasmw.exe}, and the NDISASM executable
329 files \i\c{ndisasm.exe} and \i\c{ndisasmw.exe}. In each case, the
330 file whose name ends in \c{w} is a \I{Win32}\c{Win32} executable,
331 designed to run under \I{Windows 95}\c{Windows 95} or \I{Windows NT}
332 \c{Windows NT} Intel, and the other one is a 16-bit \I{DOS}\c{DOS}
335 The only file NASM needs to run is its own executable, so copy
336 (at least) one of \c{nasm.exe} and \c{nasmw.exe} to a directory on
337 your PATH, or alternatively edit \i\c{autoexec.bat} to add the
338 \c{nasm} directory to your \i\c{PATH}. (If you're only installing the
339 \c{Win32} version, you may wish to rename it to \c{nasm.exe}.)
341 That's it - NASM is installed. You don't need the nasm directory
342 to be present to run NASM (unless you've added it to your \c{PATH}),
343 so you can delete it if you need to save space; however, you may
344 want to keep the documentation or test programs.
346 If you've downloaded the \i{DOS source archive}, \i\c{nasmXXXs.zip},
347 the \c{nasm} directory will also contain the full NASM \i{source
348 code}, and a selection of \i{Makefiles} you can (hopefully) use to
349 rebuild your copy of NASM from scratch.
351 Note that the source files \c{insnsa.c}, \c{insnsd.c}, \c{insnsi.h}
352 and \c{insnsn.c} are automatically generated from the master
353 instruction table \c{insns.dat} by a Perl script; the file
354 \c{macros.c} is generated from \c{standard.mac} by another Perl
355 script. Although the NASM source distribution includes these generated
356 files, you will need to rebuild them (and hence, will need a Perl
357 interpreter) if you change insns.dat, standard.mac or the
358 documentation. It is possible future source distributions may not
359 include these files at all. Ports of \i{Perl} for a variety of
360 platforms, including DOS and Windows, are available from
361 \W{http://www.cpan.org/ports/}\i{www.cpan.org}.
364 \S{instdos} Installing NASM under \i{Unix}
366 Once you've obtained the \i{Unix source archive} for NASM,
367 \i\c{nasm-X.XX.tar.gz} (where \c{X.XX} denotes the version number of
368 NASM contained in the archive), unpack it into a directory such
369 as \c{/usr/local/src}. The archive, when unpacked, will create its
370 own subdirectory \c{nasm-X.XX}.
372 NASM is an \I{Autoconf}\I\c{configure}auto-configuring package: once
373 you've unpacked it, \c{cd} to the directory it's been unpacked into
374 and type \c{./configure}. This shell script will find the best C
375 compiler to use for building NASM and set up \i{Makefiles}
378 Once NASM has auto-configured, you can type \i\c{make} to build the
379 \c{nasm} and \c{ndisasm} binaries, and then \c{make install} to
380 install them in \c{/usr/local/bin} and install the \i{man pages}
381 \i\c{nasm.1} and \i\c{ndisasm.1} in \c{/usr/local/man/man1}.
382 Alternatively, you can give options such as \c{--prefix} to the
383 configure script (see the file \i\c{INSTALL} for more details), or
384 install the programs yourself.
386 NASM also comes with a set of utilities for handling the \c{RDOFF}
387 custom object-file format, which are in the \i\c{rdoff} subdirectory
388 of the NASM archive. You can build these with \c{make rdf} and
389 install them with \c{make rdf_install}, if you want them.
391 If NASM fails to auto-configure, you may still be able to make it
392 compile by using the fall-back Unix makefile \i\c{Makefile.unx}.
393 Copy or rename that file to \c{Makefile} and try typing \c{make}.
394 There is also a Makefile.unx file in the \c{rdoff} subdirectory.
397 \C{running} Running NASM
399 \H{syntax} NASM \i{Command-Line} Syntax
401 To assemble a file, you issue a command of the form
403 \c nasm -f <format> <filename> [-o <output>]
407 \c nasm -f elf myfile.asm
409 will assemble \c{myfile.asm} into an \c{ELF} object file \c{myfile.o}. And
411 \c nasm -f bin myfile.asm -o myfile.com
413 will assemble \c{myfile.asm} into a raw binary file \c{myfile.com}.
415 To produce a listing file, with the hex codes output from NASM
416 displayed on the left of the original sources, use the \c{-l} option
417 to give a listing file name, for example:
419 \c nasm -f coff myfile.asm -l myfile.lst
421 To get further usage instructions from NASM, try typing
425 As \c{-hf}, this will also list the available output file formats, and what they
428 If you use Linux but aren't sure whether your system is \c{a.out}
433 (in the directory in which you put the NASM binary when you
434 installed it). If it says something like
436 \c nasm: ELF 32-bit LSB executable i386 (386 and up) Version 1
438 then your system is \c{ELF}, and you should use the option \c{-f elf}
439 when you want NASM to produce Linux object files. If it says
441 \c nasm: Linux/i386 demand-paged executable (QMAGIC)
443 or something similar, your system is \c{a.out}, and you should use
444 \c{-f aout} instead (Linux \c{a.out} systems have long been obsolete,
445 and are rare these days.)
447 Like Unix compilers and assemblers, NASM is silent unless it
448 goes wrong: you won't see any output at all, unless it gives error
452 \S{opt-o} The \i\c{-o} Option: Specifying the Output File Name
454 NASM will normally choose the name of your output file for you;
455 precisely how it does this is dependent on the object file format.
456 For Microsoft object file formats (\i\c{obj} and \i\c{win32}), it
457 will remove the \c{.asm} \i{extension} (or whatever extension you
458 like to use - NASM doesn't care) from your source file name and
459 substitute \c{.obj}. For Unix object file formats (\i\c{aout},
460 \i\c{coff}, \i\c{elf}, \i\c{macho} and \i\c{as86}) it will substitute \c{.o}. For
461 \i\c{rdf}, it will use \c{.rdf}, and for the \i\c{bin} format it
462 will simply remove the extension, so that \c{myfile.asm} produces
463 the output file \c{myfile}.
465 If the output file already exists, NASM will overwrite it, unless it
466 has the same name as the input file, in which case it will give a
467 warning and use \i\c{nasm.out} as the output file name instead.
469 For situations in which this behaviour is unacceptable, NASM
470 provides the \c{-o} command-line option, which allows you to specify
471 your desired output file name. You invoke \c{-o} by following it
472 with the name you wish for the output file, either with or without
473 an intervening space. For example:
475 \c nasm -f bin program.asm -o program.com
476 \c nasm -f bin driver.asm -odriver.sys
478 Note that this is a small o, and is different from a capital O , which
479 is used to specify the number of optimisation passes required. See \k{opt-On}.
482 \S{opt-f} The \i\c{-f} Option: Specifying the \i{Output File Format}
484 If you do not supply the \c{-f} option to NASM, it will choose an
485 output file format for you itself. In the distribution versions of
486 NASM, the default is always \i\c{bin}; if you've compiled your own
487 copy of NASM, you can redefine \i\c{OF_DEFAULT} at compile time and
488 choose what you want the default to be.
490 Like \c{-o}, the intervening space between \c{-f} and the output
491 file format is optional; so \c{-f elf} and \c{-felf} are both valid.
493 A complete list of the available output file formats can be given by
494 issuing the command \i\c{nasm -hf}.
497 \S{opt-l} The \i\c{-l} Option: Generating a \i{Listing File}
499 If you supply the \c{-l} option to NASM, followed (with the usual
500 optional space) by a file name, NASM will generate a
501 \i{source-listing file} for you, in which addresses and generated
502 code are listed on the left, and the actual source code, with
503 expansions of multi-line macros (except those which specifically
504 request no expansion in source listings: see \k{nolist}) on the
507 \c nasm -f elf myfile.asm -l myfile.lst
509 If a list file is selected, you may turn off listing for a
510 section of your source with \c{[list -]}, and turn it back on
511 with \c{[list +]}, (the default, obviously). There is no "user
512 form" (without the brackets). This can be used to list only
513 sections of interest, avoiding excessively long listings.
516 \S{opt-M} The \i\c{-M} Option: Generate \i{Makefile Dependencies}
518 This option can be used to generate makefile dependencies on stdout.
519 This can be redirected to a file for further processing. For example:
521 \c NASM -M myfile.asm > myfile.dep
524 \S{opt-MG} The \i\c{-MG} Option: Generate \i{Makefile Dependencies}
526 This option can be used to generate makefile dependencies on stdout.
527 This differs from the \c{-M} option in that if a nonexisting file is
528 encountered, it is assumed to be a generated file and is added to the
529 dependency list without a prefix.
532 \S{opt-F} The \i\c{-F} Option: Selecting a \i{Debug Information Format}
534 This option is used to select the format of the debug information emitted
535 into the output file, to be used by a debugger (or \e{will} be). Use
536 of this switch does \e{not} enable output of the selected debug info format.
537 Use \c{-g}, see \k{opt-g}, to enable output.
539 A complete list of the available debug file formats for an output format
540 can be seen by issuing the command \i\c{nasm -f <format> -y}. (only
541 "borland" in "-f obj", as of 0.98.35, but "watch this space")
544 This should not be confused with the "-f dbg" output format option which
545 is not built into NASM by default. For information on how
546 to enable it when building from the sources, see \k{dbgfmt}
549 \S{opt-g} The \i\c{-g} Option: Enabling \i{Debug Information}.
551 This option can be used to generate debugging information in the specified
552 format. See: \k{opt-F}. Using \c{-g} without \c{-F} results in emitting
553 debug info in the default format, if any, for the selected output format.
554 If no debug information is currently implemented in the selected output
555 format, \c{-g} is \e{silently ignored}.
558 \S{opt-X} The \i\c{-X} Option: Selecting an \i{Error Reporting Format}
560 This option can be used to select an error reporting format for any
561 error messages that might be produced by NASM.
563 Currently, two error reporting formats may be selected. They are
564 the \c{-Xvc} option and the \c{-Xgnu} option. The GNU format is
565 the default and looks like this:
567 \c filename.asm:65: error: specific error message
569 where \c{filename.asm} is the name of the source file in which the
570 error was detected, \c{65} is the source file line number on which
571 the error was detected, \c{error} is the severity of the error (this
572 could be \c{warning}), and \c{specific error message} is a more
573 detailed text message which should help pinpoint the exact problem.
575 The other format, specified by \c{-Xvc} is the style used by Microsoft
576 Visual C++ and some other programs. It looks like this:
578 \c filename.asm(65) : error: specific error message
580 where the only difference is that the line number is in parentheses
581 instead of being delimited by colons.
583 See also the \c{Visual C++} output format, \k{win32fmt}.
585 \S{opt-Z} The \i\c{-Z} Option: Send Errors to a File
587 Under \I{DOS}\c{MS-DOS} it can be difficult (though there are ways) to
588 redirect the standard-error output of a program to a file. Since
589 NASM usually produces its warning and \i{error messages} on
590 \i\c{stderr}, this can make it hard to capture the errors if (for
591 example) you want to load them into an editor.
593 NASM therefore provides the \c{-Z} option, taking a filename argument
594 which causes errors to be sent to the specified files rather than
595 standard error. Therefore you can \I{redirecting errors}redirect
596 the errors into a file by typing
598 \c nasm -Z myfile.err -f obj myfile.asm
600 In earlier versions of NASM, this option was called \c{-E}, but it was
601 changed since \c{-E} is an option conventionally used for
602 preprocessing only, with disastrous results. See \k{opt-E}.
604 \S{opt-s} The \i\c{-s} Option: Send Errors to \i\c{stdout}
606 The \c{-s} option redirects \i{error messages} to \c{stdout} rather
607 than \c{stderr}, so it can be redirected under \I{DOS}\c{MS-DOS}. To
608 assemble the file \c{myfile.asm} and pipe its output to the \c{more}
609 program, you can type:
611 \c nasm -s -f obj myfile.asm | more
613 See also the \c{-Z} option, \k{opt-Z}.
616 \S{opt-i} The \i\c{-i}\I\c{-I} Option: Include File Search Directories
618 When NASM sees the \i\c{%include} or \i\c{incbin} directive in
619 a source file (see \k{include} or \k{incbin}),
620 it will search for the given file not only in the
621 current directory, but also in any directories specified on the
622 command line by the use of the \c{-i} option. Therefore you can
623 include files from a \i{macro library}, for example, by typing
625 \c nasm -ic:\macrolib\ -f obj myfile.asm
627 (As usual, a space between \c{-i} and the path name is allowed, and
630 NASM, in the interests of complete source-code portability, does not
631 understand the file naming conventions of the OS it is running on;
632 the string you provide as an argument to the \c{-i} option will be
633 prepended exactly as written to the name of the include file.
634 Therefore the trailing backslash in the above example is necessary.
635 Under Unix, a trailing forward slash is similarly necessary.
637 (You can use this to your advantage, if you're really \i{perverse},
638 by noting that the option \c{-ifoo} will cause \c{%include "bar.i"}
639 to search for the file \c{foobar.i}...)
641 If you want to define a \e{standard} \i{include search path},
642 similar to \c{/usr/include} on Unix systems, you should place one or
643 more \c{-i} directives in the \c{NASMENV} environment variable (see
646 For Makefile compatibility with many C compilers, this option can also
647 be specified as \c{-I}.
650 \S{opt-p} The \i\c{-p}\I\c{-P} Option: \I{pre-including files}Pre-Include a File
652 \I\c{%include}NASM allows you to specify files to be
653 \e{pre-included} into your source file, by the use of the \c{-p}
656 \c nasm myfile.asm -p myinc.inc
658 is equivalent to running \c{nasm myfile.asm} and placing the
659 directive \c{%include "myinc.inc"} at the start of the file.
661 For consistency with the \c{-I}, \c{-D} and \c{-U} options, this
662 option can also be specified as \c{-P}.
665 \S{opt-d} The \i\c{-d}\I\c{-D} Option: \I{pre-defining macros}Pre-Define a Macro
667 \I\c{%define}Just as the \c{-p} option gives an alternative to placing
668 \c{%include} directives at the start of a source file, the \c{-d}
669 option gives an alternative to placing a \c{%define} directive. You
672 \c nasm myfile.asm -dFOO=100
674 as an alternative to placing the directive
678 at the start of the file. You can miss off the macro value, as well:
679 the option \c{-dFOO} is equivalent to coding \c{%define FOO}. This
680 form of the directive may be useful for selecting \i{assembly-time
681 options} which are then tested using \c{%ifdef}, for example
684 For Makefile compatibility with many C compilers, this option can also
685 be specified as \c{-D}.
688 \S{opt-u} The \i\c{-u}\I\c{-U} Option: \I{Undefining macros}Undefine a Macro
690 \I\c{%undef}The \c{-u} option undefines a macro that would otherwise
691 have been pre-defined, either automatically or by a \c{-p} or \c{-d}
692 option specified earlier on the command lines.
694 For example, the following command line:
696 \c nasm myfile.asm -dFOO=100 -uFOO
698 would result in \c{FOO} \e{not} being a predefined macro in the
699 program. This is useful to override options specified at a different
702 For Makefile compatibility with many C compilers, this option can also
703 be specified as \c{-U}.
706 \S{opt-E} The \i\c{-E}\I{-e} Option: Preprocess Only
708 NASM allows the \i{preprocessor} to be run on its own, up to a
709 point. Using the \c{-E} option (which requires no arguments) will
710 cause NASM to preprocess its input file, expand all the macro
711 references, remove all the comments and preprocessor directives, and
712 print the resulting file on standard output (or save it to a file,
713 if the \c{-o} option is also used).
715 This option cannot be applied to programs which require the
716 preprocessor to evaluate \I{preprocessor expressions}\i{expressions}
717 which depend on the values of symbols: so code such as
719 \c %assign tablesize ($-tablestart)
721 will cause an error in \i{preprocess-only mode}.
723 For compatiblity with older version of NASM, this option can also be
724 written \c{-e}. \c{-E} in older versions of NASM was the equivalent
725 of the current \c{-Z} option, \k{opt-Z}.
727 \S{opt-a} The \i\c{-a} Option: Don't Preprocess At All
729 If NASM is being used as the back end to a compiler, it might be
730 desirable to \I{suppressing preprocessing}suppress preprocessing
731 completely and assume the compiler has already done it, to save time
732 and increase compilation speeds. The \c{-a} option, requiring no
733 argument, instructs NASM to replace its powerful \i{preprocessor}
734 with a \i{stub preprocessor} which does nothing.
737 \S{opt-On} The \i\c{-On} Option: Specifying \i{Multipass Optimization}.
739 NASM defaults to being a two pass assembler. This means that if you
740 have a complex source file which needs more than 2 passes to assemble
741 optimally, you have to enable extra passes.
743 Using the \c{-O} option, you can tell NASM to carry out multiple passes.
746 \b \c{-O0} strict two-pass assembly, JMP and Jcc are handled more
747 like v0.98, except that backward JMPs are short, if possible.
748 Immediate operands take their long forms if a short form is
751 \b \c{-O1} strict two-pass assembly, but forward branches are assembled
752 with code guaranteed to reach; may produce larger code than
753 -O0, but will produce successful assembly more often if
754 branch offset sizes are not specified.
755 Additionally, immediate operands which will fit in a signed byte
756 are optimized, unless the long form is specified.
758 \b \c{-On} multi-pass optimization, minimize branch offsets; also will
759 minimize signed immediate bytes, overriding size specification
760 unless the \c{strict} keyword has been used (see \k{strict}).
761 The number specifies the maximum number of passes. The more
762 passes, the better the code, but the slower is the assembly.
764 Note that this is a capital O, and is different from a small o, which
765 is used to specify the output format. See \k{opt-o}.
768 \S{opt-t} The \i\c{-t} option: Enable TASM Compatibility Mode
770 NASM includes a limited form of compatibility with Borland's \i\c{TASM}.
771 When NASM's \c{-t} option is used, the following changes are made:
773 \b local labels may be prefixed with \c{@@} instead of \c{.}
775 \b TASM-style response files beginning with \c{@} may be specified on
776 the command line. This is different from the \c{-@resp} style that NASM
779 \b size override is supported within brackets. In TASM compatible mode,
780 a size override inside square brackets changes the size of the operand,
781 and not the address type of the operand as it does in NASM syntax. E.g.
782 \c{mov eax,[DWORD val]} is valid syntax in TASM compatibility mode.
783 Note that you lose the ability to override the default address type for
786 \b \c{%arg} preprocessor directive is supported which is similar to
787 TASM's \c{ARG} directive.
789 \b \c{%local} preprocessor directive
791 \b \c{%stacksize} preprocessor directive
793 \b unprefixed forms of some directives supported (\c{arg}, \c{elif},
794 \c{else}, \c{endif}, \c{if}, \c{ifdef}, \c{ifdifi}, \c{ifndef},
795 \c{include}, \c{local})
799 For more information on the directives, see the section on TASM
800 Compatiblity preprocessor directives in \k{tasmcompat}.
803 \S{opt-w} The \i\c{-w} Option: Enable or Disable Assembly \i{Warnings}
805 NASM can observe many conditions during the course of assembly which
806 are worth mentioning to the user, but not a sufficiently severe
807 error to justify NASM refusing to generate an output file. These
808 conditions are reported like errors, but come up with the word
809 `warning' before the message. Warnings do not prevent NASM from
810 generating an output file and returning a success status to the
813 Some conditions are even less severe than that: they are only
814 sometimes worth mentioning to the user. Therefore NASM supports the
815 \c{-w} command-line option, which enables or disables certain
816 classes of assembly warning. Such warning classes are described by a
817 name, for example \c{orphan-labels}; you can enable warnings of
818 this class by the command-line option \c{-w+orphan-labels} and
819 disable it by \c{-w-orphan-labels}.
821 The \i{suppressible warning} classes are:
823 \b \i\c{macro-params} covers warnings about \i{multi-line macros}
824 being invoked with the wrong number of parameters. This warning
825 class is enabled by default; see \k{mlmacover} for an example of why
826 you might want to disable it.
828 \b \i\c{macro-selfref} warns if a macro references itself. This
829 warning class is enabled by default.
831 \b \i\c{orphan-labels} covers warnings about source lines which
832 contain no instruction but define a label without a trailing colon.
833 NASM does not warn about this somewhat obscure condition by default;
834 see \k{syntax} for an example of why you might want it to.
836 \b \i\c{number-overflow} covers warnings about numeric constants which
837 don't fit in 32 bits (for example, it's easy to type one too many Fs
838 and produce \c{0x7ffffffff} by mistake). This warning class is
841 \b \i\c{gnu-elf-extensions} warns if 8-bit or 16-bit relocations
842 are used in \c{-f elf} format. The GNU extensions allow this.
843 This warning class is enabled by default.
845 \b In addition, warning classes may be enabled or disabled across
846 sections of source code with \i\c{[warning +warning-name]} or
847 \i\c{[warning -warning-name]}. No "user form" (without the
851 \S{opt-v} The \i\c{-v} Option: Display \i{Version} Info
853 Typing \c{NASM -v} will display the version of NASM which you are using,
854 and the date on which it was compiled.
856 You will need the version number if you report a bug.
858 \S{opt-y} The \i\c{-y} Option: Display Available Debug Info Formats
860 Typing \c{nasm -f <option> -y} will display a list of the available
861 debug info formats for the given output format. The default format
862 is indicated by an asterisk. E.g. \c{nasm -f obj -y} yields \c{* borland}.
863 (as of 0.98.35, the \e{only} debug info format implemented).
866 \S{opt-pfix} The \i\c{--prefix} and \i\c{--postfix} Options.
868 The \c{--prefix} and \c{--postfix} options prepend or append
869 (respectively) the given argument to all \c{global} or
870 \c{extern} variables. E.g. \c{--prefix_} will prepend the
871 underscore to all global and external variables, as C sometimes
872 (but not always) likes it.
875 \S{nasmenv} The \c{NASMENV} \i{Environment} Variable
877 If you define an environment variable called \c{NASMENV}, the program
878 will interpret it as a list of extra command-line options, which are
879 processed before the real command line. You can use this to define
880 standard search directories for include files, by putting \c{-i}
881 options in the \c{NASMENV} variable.
883 The value of the variable is split up at white space, so that the
884 value \c{-s -ic:\\nasmlib} will be treated as two separate options.
885 However, that means that the value \c{-dNAME="my name"} won't do
886 what you might want, because it will be split at the space and the
887 NASM command-line processing will get confused by the two
888 nonsensical words \c{-dNAME="my} and \c{name"}.
890 To get round this, NASM provides a feature whereby, if you begin the
891 \c{NASMENV} environment variable with some character that isn't a minus
892 sign, then NASM will treat this character as the \i{separator
893 character} for options. So setting the \c{NASMENV} variable to the
894 value \c{!-s!-ic:\\nasmlib} is equivalent to setting it to \c{-s
895 -ic:\\nasmlib}, but \c{!-dNAME="my name"} will work.
897 This environment variable was previously called \c{NASM}. This was
898 changed with version 0.98.31.
901 \H{qstart} \i{Quick Start} for \i{MASM} Users
903 If you're used to writing programs with MASM, or with \i{TASM} in
904 MASM-compatible (non-Ideal) mode, or with \i\c{a86}, this section
905 attempts to outline the major differences between MASM's syntax and
906 NASM's. If you're not already used to MASM, it's probably worth
907 skipping this section.
910 \S{qscs} NASM Is \I{case sensitivity}Case-Sensitive
912 One simple difference is that NASM is case-sensitive. It makes a
913 difference whether you call your label \c{foo}, \c{Foo} or \c{FOO}.
914 If you're assembling to \c{DOS} or \c{OS/2} \c{.OBJ} files, you can
915 invoke the \i\c{UPPERCASE} directive (documented in \k{objfmt}) to
916 ensure that all symbols exported to other code modules are forced
917 to be upper case; but even then, \e{within} a single module, NASM
918 will distinguish between labels differing only in case.
921 \S{qsbrackets} NASM Requires \i{Square Brackets} For \i{Memory References}
923 NASM was designed with simplicity of syntax in mind. One of the
924 \i{design goals} of NASM is that it should be possible, as far as is
925 practical, for the user to look at a single line of NASM code
926 and tell what opcode is generated by it. You can't do this in MASM:
927 if you declare, for example,
932 then the two lines of code
937 generate completely different opcodes, despite having
938 identical-looking syntaxes.
940 NASM avoids this undesirable situation by having a much simpler
941 syntax for memory references. The rule is simply that any access to
942 the \e{contents} of a memory location requires square brackets
943 around the address, and any access to the \e{address} of a variable
944 doesn't. So an instruction of the form \c{mov ax,foo} will
945 \e{always} refer to a compile-time constant, whether it's an \c{EQU}
946 or the address of a variable; and to access the \e{contents} of the
947 variable \c{bar}, you must code \c{mov ax,[bar]}.
949 This also means that NASM has no need for MASM's \i\c{OFFSET}
950 keyword, since the MASM code \c{mov ax,offset bar} means exactly the
951 same thing as NASM's \c{mov ax,bar}. If you're trying to get
952 large amounts of MASM code to assemble sensibly under NASM, you
953 can always code \c{%idefine offset} to make the preprocessor treat
954 the \c{OFFSET} keyword as a no-op.
956 This issue is even more confusing in \i\c{a86}, where declaring a
957 label with a trailing colon defines it to be a `label' as opposed to
958 a `variable' and causes \c{a86} to adopt NASM-style semantics; so in
959 \c{a86}, \c{mov ax,var} has different behaviour depending on whether
960 \c{var} was declared as \c{var: dw 0} (a label) or \c{var dw 0} (a
961 word-size variable). NASM is very simple by comparison:
962 \e{everything} is a label.
964 NASM, in the interests of simplicity, also does not support the
965 \i{hybrid syntaxes} supported by MASM and its clones, such as
966 \c{mov ax,table[bx]}, where a memory reference is denoted by one
967 portion outside square brackets and another portion inside. The
968 correct syntax for the above is \c{mov ax,[table+bx]}. Likewise,
969 \c{mov ax,es:[di]} is wrong and \c{mov ax,[es:di]} is right.
972 \S{qstypes} NASM Doesn't Store \i{Variable Types}
974 NASM, by design, chooses not to remember the types of variables you
975 declare. Whereas MASM will remember, on seeing \c{var dw 0}, that
976 you declared \c{var} as a word-size variable, and will then be able
977 to fill in the \i{ambiguity} in the size of the instruction \c{mov
978 var,2}, NASM will deliberately remember nothing about the symbol
979 \c{var} except where it begins, and so you must explicitly code
980 \c{mov word [var],2}.
982 For this reason, NASM doesn't support the \c{LODS}, \c{MOVS},
983 \c{STOS}, \c{SCAS}, \c{CMPS}, \c{INS}, or \c{OUTS} instructions,
984 but only supports the forms such as \c{LODSB}, \c{MOVSW}, and
985 \c{SCASD}, which explicitly specify the size of the components of
986 the strings being manipulated.
989 \S{qsassume} NASM Doesn't \i\c{ASSUME}
991 As part of NASM's drive for simplicity, it also does not support the
992 \c{ASSUME} directive. NASM will not keep track of what values you
993 choose to put in your segment registers, and will never
994 \e{automatically} generate a \i{segment override} prefix.
997 \S{qsmodel} NASM Doesn't Support \i{Memory Models}
999 NASM also does not have any directives to support different 16-bit
1000 memory models. The programmer has to keep track of which functions
1001 are supposed to be called with a \i{far call} and which with a
1002 \i{near call}, and is responsible for putting the correct form of
1003 \c{RET} instruction (\c{RETN} or \c{RETF}; NASM accepts \c{RET}
1004 itself as an alternate form for \c{RETN}); in addition, the
1005 programmer is responsible for coding CALL FAR instructions where
1006 necessary when calling \e{external} functions, and must also keep
1007 track of which external variable definitions are far and which are
1011 \S{qsfpu} \i{Floating-Point} Differences
1013 NASM uses different names to refer to floating-point registers from
1014 MASM: where MASM would call them \c{ST(0)}, \c{ST(1)} and so on, and
1015 \i\c{a86} would call them simply \c{0}, \c{1} and so on, NASM
1016 chooses to call them \c{st0}, \c{st1} etc.
1018 As of version 0.96, NASM now treats the instructions with
1019 \i{`nowait'} forms in the same way as MASM-compatible assemblers.
1020 The idiosyncratic treatment employed by 0.95 and earlier was based
1021 on a misunderstanding by the authors.
1024 \S{qsother} Other Differences
1026 For historical reasons, NASM uses the keyword \i\c{TWORD} where MASM
1027 and compatible assemblers use \i\c{TBYTE}.
1029 NASM does not declare \i{uninitialized storage} in the same way as
1030 MASM: where a MASM programmer might use \c{stack db 64 dup (?)},
1031 NASM requires \c{stack resb 64}, intended to be read as `reserve 64
1032 bytes'. For a limited amount of compatibility, since NASM treats
1033 \c{?} as a valid character in symbol names, you can code \c{? equ 0}
1034 and then writing \c{dw ?} will at least do something vaguely useful.
1035 \I\c{RESB}\i\c{DUP} is still not a supported syntax, however.
1037 In addition to all of this, macros and directives work completely
1038 differently to MASM. See \k{preproc} and \k{directive} for further
1042 \C{lang} The NASM Language
1044 \H{syntax} Layout of a NASM Source Line
1046 Like most assemblers, each NASM source line contains (unless it
1047 is a macro, a preprocessor directive or an assembler directive: see
1048 \k{preproc} and \k{directive}) some combination of the four fields
1050 \c label: instruction operands ; comment
1052 As usual, most of these fields are optional; the presence or absence
1053 of any combination of a label, an instruction and a comment is allowed.
1054 Of course, the operand field is either required or forbidden by the
1055 presence and nature of the instruction field.
1057 NASM uses backslash (\\) as the line continuation character; if a line
1058 ends with backslash, the next line is considered to be a part of the
1059 backslash-ended line.
1061 NASM places no restrictions on white space within a line: labels may
1062 have white space before them, or instructions may have no space
1063 before them, or anything. The \i{colon} after a label is also
1064 optional. (Note that this means that if you intend to code \c{lodsb}
1065 alone on a line, and type \c{lodab} by accident, then that's still a
1066 valid source line which does nothing but define a label. Running
1067 NASM with the command-line option
1068 \I{orphan-labels}\c{-w+orphan-labels} will cause it to warn you if
1069 you define a label alone on a line without a \i{trailing colon}.)
1071 \i{Valid characters} in labels are letters, numbers, \c{_}, \c{$},
1072 \c{#}, \c{@}, \c{~}, \c{.}, and \c{?}. The only characters which may
1073 be used as the \e{first} character of an identifier are letters,
1074 \c{.} (with special meaning: see \k{locallab}), \c{_} and \c{?}.
1075 An identifier may also be prefixed with a \I{$, prefix}\c{$} to
1076 indicate that it is intended to be read as an identifier and not a
1077 reserved word; thus, if some other module you are linking with
1078 defines a symbol called \c{eax}, you can refer to \c{$eax} in NASM
1079 code to distinguish the symbol from the register. Maximum length of
1080 an identifier is 4095 characters.
1082 The instruction field may contain any machine instruction: Pentium
1083 and P6 instructions, FPU instructions, MMX instructions and even
1084 undocumented instructions are all supported. The instruction may be
1085 prefixed by \c{LOCK}, \c{REP}, \c{REPE}/\c{REPZ} or
1086 \c{REPNE}/\c{REPNZ}, in the usual way. Explicit \I{address-size
1087 prefixes}address-size and \i{operand-size prefixes} \c{A16},
1088 \c{A32}, \c{O16} and \c{O32} are provided - one example of their use
1089 is given in \k{mixsize}. You can also use the name of a \I{segment
1090 override}segment register as an instruction prefix: coding
1091 \c{es mov [bx],ax} is equivalent to coding \c{mov [es:bx],ax}. We
1092 recommend the latter syntax, since it is consistent with other
1093 syntactic features of the language, but for instructions such as
1094 \c{LODSB}, which has no operands and yet can require a segment
1095 override, there is no clean syntactic way to proceed apart from
1098 An instruction is not required to use a prefix: prefixes such as
1099 \c{CS}, \c{A32}, \c{LOCK} or \c{REPE} can appear on a line by
1100 themselves, and NASM will just generate the prefix bytes.
1102 In addition to actual machine instructions, NASM also supports a
1103 number of pseudo-instructions, described in \k{pseudop}.
1105 Instruction \i{operands} may take a number of forms: they can be
1106 registers, described simply by the register name (e.g. \c{ax},
1107 \c{bp}, \c{ebx}, \c{cr0}: NASM does not use the \c{gas}-style
1108 syntax in which register names must be prefixed by a \c{%} sign), or
1109 they can be \i{effective addresses} (see \k{effaddr}), constants
1110 (\k{const}) or expressions (\k{expr}).
1112 For x87 \i{floating-point} instructions, NASM accepts a wide range of
1113 syntaxes: you can use two-operand forms like MASM supports, or you
1114 can use NASM's native single-operand forms in most cases.
1116 \# all forms of each supported instruction are given in
1118 For example, you can code:
1120 \c fadd st1 ; this sets st0 := st0 + st1
1121 \c fadd st0,st1 ; so does this
1123 \c fadd st1,st0 ; this sets st1 := st1 + st0
1124 \c fadd to st1 ; so does this
1126 Almost any x87 floating-point instruction that references memory must
1127 use one of the prefixes \i\c{DWORD}, \i\c{QWORD} or \i\c{TWORD} to
1128 indicate what size of \i{memory operand} it refers to.
1131 \H{pseudop} \i{Pseudo-Instructions}
1133 Pseudo-instructions are things which, though not real x86 machine
1134 instructions, are used in the instruction field anyway because that's
1135 the most convenient place to put them. The current pseudo-instructions
1136 are \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT} and \i\c{DO};
1137 their \i{uninitialized} counterparts \i\c{RESB}, \i\c{RESW},
1138 \i\c{RESD}, \i\c{RESQ}, \i\c{REST} and \i\c{RESO}; the \i\c{INCBIN}
1139 command, the \i\c{EQU} command, and the \i\c{TIMES} prefix.
1142 \S{db} \c{DB} and friends: Declaring initialized Data
1144 \i\c{DB}, \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT} and \i\c{DO} are
1145 used, much as in MASM, to declare initialized data in the output
1146 file. They can be invoked in a wide range of ways:
1147 \I{floating-point}\I{character constant}\I{string constant}
1149 \c db 0x55 ; just the byte 0x55
1150 \c db 0x55,0x56,0x57 ; three bytes in succession
1151 \c db 'a',0x55 ; character constants are OK
1152 \c db 'hello',13,10,'$' ; so are string constants
1153 \c dw 0x1234 ; 0x34 0x12
1154 \c dw 'a' ; 0x61 0x00 (it's just a number)
1155 \c dw 'ab' ; 0x61 0x62 (character constant)
1156 \c dw 'abc' ; 0x61 0x62 0x63 0x00 (string)
1157 \c dd 0x12345678 ; 0x78 0x56 0x34 0x12
1158 \c dd 1.234567e20 ; floating-point constant
1159 \c dq 0x123456789abcdef0 ; eight byte constant
1160 \c dq 1.234567e20 ; double-precision float
1161 \c dt 1.234567e20 ; extended-precision float
1163 \c{DT} and \c{DO} do not accept \i{numeric constants} as operands.
1164 \c{DB} does not accept \i{floating-point} numbers as operands.
1167 \S{resb} \c{RESB} and friends: Declaring \i{Uninitialized} Data
1169 \i\c{RESB}, \i\c{RESW}, \i\c{RESD}, \i\c{RESQ}, \i\c{REST} and
1170 \i\c{RESO} are designed to be used in the BSS section of a module:
1171 they declare \e{uninitialized} storage space. Each takes a single
1172 operand, which is the number of bytes, words, doublewords or whatever
1173 to reserve. As stated in \k{qsother}, NASM does not support the
1174 MASM/TASM syntax of reserving uninitialized space by writing
1175 \I\c{?}\c{DW ?} or similar things: this is what it does instead. The
1176 operand to a \c{RESB}-type pseudo-instruction is a \i\e{critical
1177 expression}: see \k{crit}.
1181 \c buffer: resb 64 ; reserve 64 bytes
1182 \c wordvar: resw 1 ; reserve a word
1183 \c realarray resq 10 ; array of ten reals
1186 \S{incbin} \i\c{INCBIN}: Including External \i{Binary Files}
1188 \c{INCBIN} is borrowed from the old Amiga assembler \i{DevPac}: it
1189 includes a binary file verbatim into the output file. This can be
1190 handy for (for example) including \i{graphics} and \i{sound} data
1191 directly into a game executable file. It can be called in one of
1194 \c incbin "file.dat" ; include the whole file
1195 \c incbin "file.dat",1024 ; skip the first 1024 bytes
1196 \c incbin "file.dat",1024,512 ; skip the first 1024, and
1197 \c ; actually include at most 512
1200 \S{equ} \i\c{EQU}: Defining Constants
1202 \c{EQU} defines a symbol to a given constant value: when \c{EQU} is
1203 used, the source line must contain a label. The action of \c{EQU} is
1204 to define the given label name to the value of its (only) operand.
1205 This definition is absolute, and cannot change later. So, for
1208 \c message db 'hello, world'
1209 \c msglen equ $-message
1211 defines \c{msglen} to be the constant 12. \c{msglen} may not then be
1212 redefined later. This is not a \i{preprocessor} definition either:
1213 the value of \c{msglen} is evaluated \e{once}, using the value of
1214 \c{$} (see \k{expr} for an explanation of \c{$}) at the point of
1215 definition, rather than being evaluated wherever it is referenced
1216 and using the value of \c{$} at the point of reference. Note that
1217 the operand to an \c{EQU} is also a \i{critical expression}
1221 \S{times} \i\c{TIMES}: \i{Repeating} Instructions or Data
1223 The \c{TIMES} prefix causes the instruction to be assembled multiple
1224 times. This is partly present as NASM's equivalent of the \i\c{DUP}
1225 syntax supported by \i{MASM}-compatible assemblers, in that you can
1228 \c zerobuf: times 64 db 0
1230 or similar things; but \c{TIMES} is more versatile than that. The
1231 argument to \c{TIMES} is not just a numeric constant, but a numeric
1232 \e{expression}, so you can do things like
1234 \c buffer: db 'hello, world'
1235 \c times 64-$+buffer db ' '
1237 which will store exactly enough spaces to make the total length of
1238 \c{buffer} up to 64. Finally, \c{TIMES} can be applied to ordinary
1239 instructions, so you can code trivial \i{unrolled loops} in it:
1243 Note that there is no effective difference between \c{times 100 resb
1244 1} and \c{resb 100}, except that the latter will be assembled about
1245 100 times faster due to the internal structure of the assembler.
1247 The operand to \c{TIMES}, like that of \c{EQU} and those of \c{RESB}
1248 and friends, is a critical expression (\k{crit}).
1250 Note also that \c{TIMES} can't be applied to \i{macros}: the reason
1251 for this is that \c{TIMES} is processed after the macro phase, which
1252 allows the argument to \c{TIMES} to contain expressions such as
1253 \c{64-$+buffer} as above. To repeat more than one line of code, or a
1254 complex macro, use the preprocessor \i\c{%rep} directive.
1257 \H{effaddr} Effective Addresses
1259 An \i{effective address} is any operand to an instruction which
1260 \I{memory reference}references memory. Effective addresses, in NASM,
1261 have a very simple syntax: they consist of an expression evaluating
1262 to the desired address, enclosed in \i{square brackets}. For
1267 \c mov ax,[wordvar+1]
1268 \c mov ax,[es:wordvar+bx]
1270 Anything not conforming to this simple system is not a valid memory
1271 reference in NASM, for example \c{es:wordvar[bx]}.
1273 More complicated effective addresses, such as those involving more
1274 than one register, work in exactly the same way:
1276 \c mov eax,[ebx*2+ecx+offset]
1279 NASM is capable of doing \i{algebra} on these effective addresses,
1280 so that things which don't necessarily \e{look} legal are perfectly
1283 \c mov eax,[ebx*5] ; assembles as [ebx*4+ebx]
1284 \c mov eax,[label1*2-label2] ; ie [label1+(label1-label2)]
1286 Some forms of effective address have more than one assembled form;
1287 in most such cases NASM will generate the smallest form it can. For
1288 example, there are distinct assembled forms for the 32-bit effective
1289 addresses \c{[eax*2+0]} and \c{[eax+eax]}, and NASM will generally
1290 generate the latter on the grounds that the former requires four
1291 bytes to store a zero offset.
1293 NASM has a hinting mechanism which will cause \c{[eax+ebx]} and
1294 \c{[ebx+eax]} to generate different opcodes; this is occasionally
1295 useful because \c{[esi+ebp]} and \c{[ebp+esi]} have different
1296 default segment registers.
1298 However, you can force NASM to generate an effective address in a
1299 particular form by the use of the keywords \c{BYTE}, \c{WORD},
1300 \c{DWORD} and \c{NOSPLIT}. If you need \c{[eax+3]} to be assembled
1301 using a double-word offset field instead of the one byte NASM will
1302 normally generate, you can code \c{[dword eax+3]}. Similarly, you
1303 can force NASM to use a byte offset for a small value which it
1304 hasn't seen on the first pass (see \k{crit} for an example of such a
1305 code fragment) by using \c{[byte eax+offset]}. As special cases,
1306 \c{[byte eax]} will code \c{[eax+0]} with a byte offset of zero, and
1307 \c{[dword eax]} will code it with a double-word offset of zero. The
1308 normal form, \c{[eax]}, will be coded with no offset field.
1310 The form described in the previous paragraph is also useful if you
1311 are trying to access data in a 32-bit segment from within 16 bit code.
1312 For more information on this see the section on mixed-size addressing
1313 (\k{mixaddr}). In particular, if you need to access data with a known
1314 offset that is larger than will fit in a 16-bit value, if you don't
1315 specify that it is a dword offset, nasm will cause the high word of
1316 the offset to be lost.
1318 Similarly, NASM will split \c{[eax*2]} into \c{[eax+eax]} because
1319 that allows the offset field to be absent and space to be saved; in
1320 fact, it will also split \c{[eax*2+offset]} into
1321 \c{[eax+eax+offset]}. You can combat this behaviour by the use of
1322 the \c{NOSPLIT} keyword: \c{[nosplit eax*2]} will force
1323 \c{[eax*2+0]} to be generated literally.
1325 In 64-bit mode, NASM will by default generate absolute addresses. The
1326 \i\c{REL} keyword makes it produce \c{RIP}-relative addresses. Since
1327 this is frequently the normally desired behaviour, see the \c{DEFAULT}
1328 directive (\k{default}). The keyword \i\c{ABS} overrides \i\c{REL}.
1331 \H{const} \i{Constants}
1333 NASM understands four different types of constant: numeric,
1334 character, string and floating-point.
1337 \S{numconst} \i{Numeric Constants}
1339 A numeric constant is simply a number. NASM allows you to specify
1340 numbers in a variety of number bases, in a variety of ways: you can
1341 suffix \c{H}, \c{Q} or \c{O}, and \c{B} for \i{hex}, \i{octal} and \i{binary},
1342 or you can prefix \c{0x} for hex in the style of C, or you can
1343 prefix \c{$} for hex in the style of Borland Pascal. Note, though,
1344 that the \I{$, prefix}\c{$} prefix does double duty as a prefix on
1345 identifiers (see \k{syntax}), so a hex number prefixed with a \c{$}
1346 sign must have a digit after the \c{$} rather than a letter.
1350 \c mov ax,100 ; decimal
1351 \c mov ax,0a2h ; hex
1352 \c mov ax,$0a2 ; hex again: the 0 is required
1353 \c mov ax,0xa2 ; hex yet again
1354 \c mov ax,777q ; octal
1355 \c mov ax,777o ; octal again
1356 \c mov ax,10010011b ; binary
1359 \S{chrconst} \i{Character Constants}
1361 A character constant consists of up to four characters enclosed in
1362 either single or double quotes. The type of quote makes no
1363 difference to NASM, except of course that surrounding the constant
1364 with single quotes allows double quotes to appear within it and vice
1367 A character constant with more than one character will be arranged
1368 with \i{little-endian} order in mind: if you code
1372 then the constant generated is not \c{0x61626364}, but
1373 \c{0x64636261}, so that if you were then to store the value into
1374 memory, it would read \c{abcd} rather than \c{dcba}. This is also
1375 the sense of character constants understood by the Pentium's
1376 \i\c{CPUID} instruction.
1377 \# (see \k{insCPUID})
1380 \S{strconst} String Constants
1382 String constants are only acceptable to some pseudo-instructions,
1383 namely the \I\c{DW}\I\c{DD}\I\c{DQ}\I\c{DT}\i\c{DB} family and
1386 A string constant looks like a character constant, only longer. It
1387 is treated as a concatenation of maximum-size character constants
1388 for the conditions. So the following are equivalent:
1390 \c db 'hello' ; string constant
1391 \c db 'h','e','l','l','o' ; equivalent character constants
1393 And the following are also equivalent:
1395 \c dd 'ninechars' ; doubleword string constant
1396 \c dd 'nine','char','s' ; becomes three doublewords
1397 \c db 'ninechars',0,0,0 ; and really looks like this
1399 Note that when used as an operand to \c{db}, a constant like
1400 \c{'ab'} is treated as a string constant despite being short enough
1401 to be a character constant, because otherwise \c{db 'ab'} would have
1402 the same effect as \c{db 'a'}, which would be silly. Similarly,
1403 three-character or four-character constants are treated as strings
1404 when they are operands to \c{dw}.
1407 \S{fltconst} \I{floating-point, constants}Floating-Point Constants
1409 \i{Floating-point} constants are acceptable only as arguments to
1410 \i\c{DW}, \i\c{DD}, \i\c{DQ}, \i\c{DT}, and \i\c{DO}, or as arguments
1411 to the special operators \i\c{__float16__}, \i\c{__float32__},
1412 \i\c{__float64__}, \i\c{__float80m__}, \i\c{__float80e__},
1413 \i\c{__float128l__}, and \i\c{__float128h__}.
1415 Floating-point constants are expressed in the traditional form:
1416 digits, then a period, then optionally more digits, then optionally an
1417 \c{E} followed by an exponent. The period is mandatory, so that NASM
1418 can distinguish between \c{dd 1}, which declares an integer constant,
1419 and \c{dd 1.0} which declares a floating-point constant. NASM also
1420 support C99-style hexadecimal floating-point: \c{0x}, hexadecimal
1421 digits, period, optionally more hexadeximal digits, then optionally a
1422 \c{P} followed by a \e{binary} (not hexadecimal) exponent in decimal
1427 \c dw -0.5 ; IEEE half precision
1428 \c dd 1.2 ; an easy one
1429 \c dd 0x1p+2 ; 1.0x2^2 = 4.0
1430 \c dq 1.e10 ; 10,000,000,000
1431 \c dq 1.e+10 ; synonymous with 1.e10
1432 \c dq 1.e-10 ; 0.000 000 000 1
1433 \c dt 3.141592653589793238462 ; pi
1434 \c do 1.e+4000 ; IEEE quad precision
1436 The special operators are used to produce floating-point numbers in
1437 other contexts. They produce the binary representation of a specific
1438 floating-point number as an integer, and can use anywhere integer
1439 constants are used in an expression. \c{__float80m__} and
1440 \c{__float80e__} produce the 64-bit mantissa and 16-bit exponent of an
1441 80-bit floating-point number, and \c{__float128l__} and
1442 \c{__float128h__} produce the lower and upper 64-bit halves of a 128-bit
1443 floating-point number, respectively.
1447 \c mov rax,__float64__(3.141592653589793238462)
1449 ... would assign the binary representation of pi as a 64-bit floating
1450 point number into \c{RAX}. This is exactly equivalent to:
1452 \c mov rax,0x401921fb54442d18
1454 NASM cannot do compile-time arithmetic on floating-point constants.
1455 This is because NASM is designed to be portable - although it always
1456 generates code to run on x86 processors, the assembler itself can
1457 run on any system with an ANSI C compiler. Therefore, the assembler
1458 cannot guarantee the presence of a floating-point unit capable of
1459 handling the \i{Intel number formats}, and so for NASM to be able to
1460 do floating arithmetic it would have to include its own complete set
1461 of floating-point routines, which would significantly increase the
1462 size of the assembler for very little benefit.
1464 The special tokens \i\c{__Infinity__}, \i\c{__QNaN__} (or
1465 \i\c{__NaN__}) and \i\c{__SNaN__} can be used to generate
1466 \I{infinity}infinities, quiet \i{NaN}s, and signalling NaNs,
1467 respectively. These are normally used as macros:
1469 \c %define Inf __Infinity__
1470 \c %define NaN __QNaN__
1472 \c dq +1.5, -Inf, NaN ; Double-precision constants
1474 \H{expr} \i{Expressions}
1476 Expressions in NASM are similar in syntax to those in C. Expressions
1477 are evaluated as 64-bit integers which are then adjusted to the
1480 NASM supports two special tokens in expressions, allowing
1481 calculations to involve the current assembly position: the
1482 \I{$, here}\c{$} and \i\c{$$} tokens. \c{$} evaluates to the assembly
1483 position at the beginning of the line containing the expression; so
1484 you can code an \i{infinite loop} using \c{JMP $}. \c{$$} evaluates
1485 to the beginning of the current section; so you can tell how far
1486 into the section you are by using \c{($-$$)}.
1488 The arithmetic \i{operators} provided by NASM are listed here, in
1489 increasing order of \i{precedence}.
1492 \S{expor} \i\c{|}: \i{Bitwise OR} Operator
1494 The \c{|} operator gives a bitwise OR, exactly as performed by the
1495 \c{OR} machine instruction. Bitwise OR is the lowest-priority
1496 arithmetic operator supported by NASM.
1499 \S{expxor} \i\c{^}: \i{Bitwise XOR} Operator
1501 \c{^} provides the bitwise XOR operation.
1504 \S{expand} \i\c{&}: \i{Bitwise AND} Operator
1506 \c{&} provides the bitwise AND operation.
1509 \S{expshift} \i\c{<<} and \i\c{>>}: \i{Bit Shift} Operators
1511 \c{<<} gives a bit-shift to the left, just as it does in C. So \c{5<<3}
1512 evaluates to 5 times 8, or 40. \c{>>} gives a bit-shift to the
1513 right; in NASM, such a shift is \e{always} unsigned, so that
1514 the bits shifted in from the left-hand end are filled with zero
1515 rather than a sign-extension of the previous highest bit.
1518 \S{expplmi} \I{+ opaddition}\c{+} and \I{- opsubtraction}\c{-}:
1519 \i{Addition} and \i{Subtraction} Operators
1521 The \c{+} and \c{-} operators do perfectly ordinary addition and
1525 \S{expmul} \i\c{*}, \i\c{/}, \i\c{//}, \i\c{%} and \i\c{%%}:
1526 \i{Multiplication} and \i{Division}
1528 \c{*} is the multiplication operator. \c{/} and \c{//} are both
1529 division operators: \c{/} is \i{unsigned division} and \c{//} is
1530 \i{signed division}. Similarly, \c{%} and \c{%%} provide \I{unsigned
1531 modulo}\I{modulo operators}unsigned and
1532 \i{signed modulo} operators respectively.
1534 NASM, like ANSI C, provides no guarantees about the sensible
1535 operation of the signed modulo operator.
1537 Since the \c{%} character is used extensively by the macro
1538 \i{preprocessor}, you should ensure that both the signed and unsigned
1539 modulo operators are followed by white space wherever they appear.
1542 \S{expmul} \i{Unary Operators}: \I{+ opunary}\c{+}, \I{- opunary}\c{-},
1543 \i\c{~}, \I{! opunary}\c{!} and \i\c{SEG}
1545 The highest-priority operators in NASM's expression grammar are
1546 those which only apply to one argument. \c{-} negates its operand,
1547 \c{+} does nothing (it's provided for symmetry with \c{-}), \c{~}
1548 computes the \i{one's complement} of its operand, \c{!} is the
1549 \i{logical negation} operator, and \c{SEG} provides the \i{segment address}
1550 of its operand (explained in more detail in \k{segwrt}).
1553 \H{segwrt} \i\c{SEG} and \i\c{WRT}
1555 When writing large 16-bit programs, which must be split into
1556 multiple \i{segments}, it is often necessary to be able to refer to
1557 the \I{segment address}segment part of the address of a symbol. NASM
1558 supports the \c{SEG} operator to perform this function.
1560 The \c{SEG} operator returns the \i\e{preferred} segment base of a
1561 symbol, defined as the segment base relative to which the offset of
1562 the symbol makes sense. So the code
1564 \c mov ax,seg symbol
1568 will load \c{ES:BX} with a valid pointer to the symbol \c{symbol}.
1570 Things can be more complex than this: since 16-bit segments and
1571 \i{groups} may \I{overlapping segments}overlap, you might occasionally
1572 want to refer to some symbol using a different segment base from the
1573 preferred one. NASM lets you do this, by the use of the \c{WRT}
1574 (With Reference To) keyword. So you can do things like
1576 \c mov ax,weird_seg ; weird_seg is a segment base
1578 \c mov bx,symbol wrt weird_seg
1580 to load \c{ES:BX} with a different, but functionally equivalent,
1581 pointer to the symbol \c{symbol}.
1583 NASM supports far (inter-segment) calls and jumps by means of the
1584 syntax \c{call segment:offset}, where \c{segment} and \c{offset}
1585 both represent immediate values. So to call a far procedure, you
1586 could code either of
1588 \c call (seg procedure):procedure
1589 \c call weird_seg:(procedure wrt weird_seg)
1591 (The parentheses are included for clarity, to show the intended
1592 parsing of the above instructions. They are not necessary in
1595 NASM supports the syntax \I\c{CALL FAR}\c{call far procedure} as a
1596 synonym for the first of the above usages. \c{JMP} works identically
1597 to \c{CALL} in these examples.
1599 To declare a \i{far pointer} to a data item in a data segment, you
1602 \c dw symbol, seg symbol
1604 NASM supports no convenient synonym for this, though you can always
1605 invent one using the macro processor.
1608 \H{strict} \i\c{STRICT}: Inhibiting Optimization
1610 When assembling with the optimizer set to level 2 or higher (see
1611 \k{opt-On}), NASM will use size specifiers (\c{BYTE}, \c{WORD},
1612 \c{DWORD}, \c{QWORD}, \c{TWORD} or \c{OWORD}), but will give them the
1613 smallest possible size. The keyword \c{STRICT} can be used to inhibit
1614 optimization and force a particular operand to be emitted in the
1615 specified size. For example, with the optimizer on, and in \c{BITS 16}
1620 is encoded in three bytes \c{66 6A 21}, whereas
1622 \c push strict dword 33
1624 is encoded in six bytes, with a full dword immediate operand \c{66 68
1627 With the optimizer off, the same code (six bytes) is generated whether
1628 the \c{STRICT} keyword was used or not.
1631 \H{crit} \i{Critical Expressions}
1633 A limitation of NASM is that it is a \i{two-pass assembler}; unlike
1634 TASM and others, it will always do exactly two \I{passes}\i{assembly
1635 passes}. Therefore it is unable to cope with source files that are
1636 complex enough to require three or more passes.
1638 The first pass is used to determine the size of all the assembled
1639 code and data, so that the second pass, when generating all the
1640 code, knows all the symbol addresses the code refers to. So one
1641 thing NASM can't handle is code whose size depends on the value of a
1642 symbol declared after the code in question. For example,
1644 \c times (label-$) db 0
1645 \c label: db 'Where am I?'
1647 The argument to \i\c{TIMES} in this case could equally legally
1648 evaluate to anything at all; NASM will reject this example because
1649 it cannot tell the size of the \c{TIMES} line when it first sees it.
1650 It will just as firmly reject the slightly \I{paradox}paradoxical
1653 \c times (label-$+1) db 0
1654 \c label: db 'NOW where am I?'
1656 in which \e{any} value for the \c{TIMES} argument is by definition
1659 NASM rejects these examples by means of a concept called a
1660 \e{critical expression}, which is defined to be an expression whose
1661 value is required to be computable in the first pass, and which must
1662 therefore depend only on symbols defined before it. The argument to
1663 the \c{TIMES} prefix is a critical expression; for the same reason,
1664 the arguments to the \i\c{RESB} family of pseudo-instructions are
1665 also critical expressions.
1667 Critical expressions can crop up in other contexts as well: consider
1671 \c symbol1 equ symbol2
1674 On the first pass, NASM cannot determine the value of \c{symbol1},
1675 because \c{symbol1} is defined to be equal to \c{symbol2} which NASM
1676 hasn't seen yet. On the second pass, therefore, when it encounters
1677 the line \c{mov ax,symbol1}, it is unable to generate the code for
1678 it because it still doesn't know the value of \c{symbol1}. On the
1679 next line, it would see the \i\c{EQU} again and be able to determine
1680 the value of \c{symbol1}, but by then it would be too late.
1682 NASM avoids this problem by defining the right-hand side of an
1683 \c{EQU} statement to be a critical expression, so the definition of
1684 \c{symbol1} would be rejected in the first pass.
1686 There is a related issue involving \i{forward references}: consider
1689 \c mov eax,[ebx+offset]
1692 NASM, on pass one, must calculate the size of the instruction \c{mov
1693 eax,[ebx+offset]} without knowing the value of \c{offset}. It has no
1694 way of knowing that \c{offset} is small enough to fit into a
1695 one-byte offset field and that it could therefore get away with
1696 generating a shorter form of the \i{effective-address} encoding; for
1697 all it knows, in pass one, \c{offset} could be a symbol in the code
1698 segment, and it might need the full four-byte form. So it is forced
1699 to compute the size of the instruction to accommodate a four-byte
1700 address part. In pass two, having made this decision, it is now
1701 forced to honour it and keep the instruction large, so the code
1702 generated in this case is not as small as it could have been. This
1703 problem can be solved by defining \c{offset} before using it, or by
1704 forcing byte size in the effective address by coding \c{[byte
1707 Note that use of the \c{-On} switch (with n>=2) makes some of the above
1708 no longer true (see \k{opt-On}).
1710 \H{locallab} \i{Local Labels}
1712 NASM gives special treatment to symbols beginning with a \i{period}.
1713 A label beginning with a single period is treated as a \e{local}
1714 label, which means that it is associated with the previous non-local
1715 label. So, for example:
1717 \c label1 ; some code
1725 \c label2 ; some code
1733 In the above code fragment, each \c{JNE} instruction jumps to the
1734 line immediately before it, because the two definitions of \c{.loop}
1735 are kept separate by virtue of each being associated with the
1736 previous non-local label.
1738 This form of local label handling is borrowed from the old Amiga
1739 assembler \i{DevPac}; however, NASM goes one step further, in
1740 allowing access to local labels from other parts of the code. This
1741 is achieved by means of \e{defining} a local label in terms of the
1742 previous non-local label: the first definition of \c{.loop} above is
1743 really defining a symbol called \c{label1.loop}, and the second
1744 defines a symbol called \c{label2.loop}. So, if you really needed
1747 \c label3 ; some more code
1752 Sometimes it is useful - in a macro, for instance - to be able to
1753 define a label which can be referenced from anywhere but which
1754 doesn't interfere with the normal local-label mechanism. Such a
1755 label can't be non-local because it would interfere with subsequent
1756 definitions of, and references to, local labels; and it can't be
1757 local because the macro that defined it wouldn't know the label's
1758 full name. NASM therefore introduces a third type of label, which is
1759 probably only useful in macro definitions: if a label begins with
1760 the \I{label prefix}special prefix \i\c{..@}, then it does nothing
1761 to the local label mechanism. So you could code
1763 \c label1: ; a non-local label
1764 \c .local: ; this is really label1.local
1765 \c ..@foo: ; this is a special symbol
1766 \c label2: ; another non-local label
1767 \c .local: ; this is really label2.local
1769 \c jmp ..@foo ; this will jump three lines up
1771 NASM has the capacity to define other special symbols beginning with
1772 a double period: for example, \c{..start} is used to specify the
1773 entry point in the \c{obj} output format (see \k{dotdotstart}).
1776 \C{preproc} The NASM \i{Preprocessor}
1778 NASM contains a powerful \i{macro processor}, which supports
1779 conditional assembly, multi-level file inclusion, two forms of macro
1780 (single-line and multi-line), and a `context stack' mechanism for
1781 extra macro power. Preprocessor directives all begin with a \c{%}
1784 The preprocessor collapses all lines which end with a backslash (\\)
1785 character into a single line. Thus:
1787 \c %define THIS_VERY_LONG_MACRO_NAME_IS_DEFINED_TO \\
1790 will work like a single-line macro without the backslash-newline
1793 \H{slmacro} \i{Single-Line Macros}
1795 \S{define} The Normal Way: \I\c{%idefine}\i\c{%define}
1797 Single-line macros are defined using the \c{%define} preprocessor
1798 directive. The definitions work in a similar way to C; so you can do
1801 \c %define ctrl 0x1F &
1802 \c %define param(a,b) ((a)+(a)*(b))
1804 \c mov byte [param(2,ebx)], ctrl 'D'
1806 which will expand to
1808 \c mov byte [(2)+(2)*(ebx)], 0x1F & 'D'
1810 When the expansion of a single-line macro contains tokens which
1811 invoke another macro, the expansion is performed at invocation time,
1812 not at definition time. Thus the code
1814 \c %define a(x) 1+b(x)
1819 will evaluate in the expected way to \c{mov ax,1+2*8}, even though
1820 the macro \c{b} wasn't defined at the time of definition of \c{a}.
1822 Macros defined with \c{%define} are \i{case sensitive}: after
1823 \c{%define foo bar}, only \c{foo} will expand to \c{bar}: \c{Foo} or
1824 \c{FOO} will not. By using \c{%idefine} instead of \c{%define} (the
1825 `i' stands for `insensitive') you can define all the case variants
1826 of a macro at once, so that \c{%idefine foo bar} would cause
1827 \c{foo}, \c{Foo}, \c{FOO}, \c{fOO} and so on all to expand to
1830 There is a mechanism which detects when a macro call has occurred as
1831 a result of a previous expansion of the same macro, to guard against
1832 \i{circular references} and infinite loops. If this happens, the
1833 preprocessor will only expand the first occurrence of the macro.
1836 \c %define a(x) 1+a(x)
1840 the macro \c{a(3)} will expand once, becoming \c{1+a(3)}, and will
1841 then expand no further. This behaviour can be useful: see \k{32c}
1842 for an example of its use.
1844 You can \I{overloading, single-line macros}overload single-line
1845 macros: if you write
1847 \c %define foo(x) 1+x
1848 \c %define foo(x,y) 1+x*y
1850 the preprocessor will be able to handle both types of macro call,
1851 by counting the parameters you pass; so \c{foo(3)} will become
1852 \c{1+3} whereas \c{foo(ebx,2)} will become \c{1+ebx*2}. However, if
1857 then no other definition of \c{foo} will be accepted: a macro with
1858 no parameters prohibits the definition of the same name as a macro
1859 \e{with} parameters, and vice versa.
1861 This doesn't prevent single-line macros being \e{redefined}: you can
1862 perfectly well define a macro with
1866 and then re-define it later in the same source file with
1870 Then everywhere the macro \c{foo} is invoked, it will be expanded
1871 according to the most recent definition. This is particularly useful
1872 when defining single-line macros with \c{%assign} (see \k{assign}).
1874 You can \i{pre-define} single-line macros using the `-d' option on
1875 the NASM command line: see \k{opt-d}.
1878 \S{xdefine} Enhancing %define: \I\c{%xidefine}\i\c{%xdefine}
1880 To have a reference to an embedded single-line macro resolved at the
1881 time that it is embedded, as opposed to when the calling macro is
1882 expanded, you need a different mechanism to the one offered by
1883 \c{%define}. The solution is to use \c{%xdefine}, or it's
1884 \I{case sensitive}case-insensitive counterpart \c{%xidefine}.
1886 Suppose you have the following code:
1889 \c %define isFalse isTrue
1898 In this case, \c{val1} is equal to 0, and \c{val2} is equal to 1.
1899 This is because, when a single-line macro is defined using
1900 \c{%define}, it is expanded only when it is called. As \c{isFalse}
1901 expands to \c{isTrue}, the expansion will be the current value of
1902 \c{isTrue}. The first time it is called that is 0, and the second
1905 If you wanted \c{isFalse} to expand to the value assigned to the
1906 embedded macro \c{isTrue} at the time that \c{isFalse} was defined,
1907 you need to change the above code to use \c{%xdefine}.
1909 \c %xdefine isTrue 1
1910 \c %xdefine isFalse isTrue
1911 \c %xdefine isTrue 0
1915 \c %xdefine isTrue 1
1919 Now, each time that \c{isFalse} is called, it expands to 1,
1920 as that is what the embedded macro \c{isTrue} expanded to at
1921 the time that \c{isFalse} was defined.
1924 \S{concat%+} Concatenating Single Line Macro Tokens: \i\c{%+}
1926 Individual tokens in single line macros can be concatenated, to produce
1927 longer tokens for later processing. This can be useful if there are
1928 several similar macros that perform similar functions.
1930 Please note that a space is required after \c{%+}, in order to
1931 disambiguate it from the syntax \c{%+1} used in multiline macros.
1933 As an example, consider the following:
1935 \c %define BDASTART 400h ; Start of BIOS data area
1937 \c struc tBIOSDA ; its structure
1943 Now, if we need to access the elements of tBIOSDA in different places,
1946 \c mov ax,BDASTART + tBIOSDA.COM1addr
1947 \c mov bx,BDASTART + tBIOSDA.COM2addr
1949 This will become pretty ugly (and tedious) if used in many places, and
1950 can be reduced in size significantly by using the following macro:
1952 \c ; Macro to access BIOS variables by their names (from tBDA):
1954 \c %define BDA(x) BDASTART + tBIOSDA. %+ x
1956 Now the above code can be written as:
1958 \c mov ax,BDA(COM1addr)
1959 \c mov bx,BDA(COM2addr)
1961 Using this feature, we can simplify references to a lot of macros (and,
1962 in turn, reduce typing errors).
1965 \S{undef} Undefining macros: \i\c{%undef}
1967 Single-line macros can be removed with the \c{%undef} command. For
1968 example, the following sequence:
1975 will expand to the instruction \c{mov eax, foo}, since after
1976 \c{%undef} the macro \c{foo} is no longer defined.
1978 Macros that would otherwise be pre-defined can be undefined on the
1979 command-line using the `-u' option on the NASM command line: see
1983 \S{assign} \i{Preprocessor Variables}: \i\c{%assign}
1985 An alternative way to define single-line macros is by means of the
1986 \c{%assign} command (and its \I{case sensitive}case-insensitive
1987 counterpart \i\c{%iassign}, which differs from \c{%assign} in
1988 exactly the same way that \c{%idefine} differs from \c{%define}).
1990 \c{%assign} is used to define single-line macros which take no
1991 parameters and have a numeric value. This value can be specified in
1992 the form of an expression, and it will be evaluated once, when the
1993 \c{%assign} directive is processed.
1995 Like \c{%define}, macros defined using \c{%assign} can be re-defined
1996 later, so you can do things like
2000 to increment the numeric value of a macro.
2002 \c{%assign} is useful for controlling the termination of \c{%rep}
2003 preprocessor loops: see \k{rep} for an example of this. Another
2004 use for \c{%assign} is given in \k{16c} and \k{32c}.
2006 The expression passed to \c{%assign} is a \i{critical expression}
2007 (see \k{crit}), and must also evaluate to a pure number (rather than
2008 a relocatable reference such as a code or data address, or anything
2009 involving a register).
2012 \H{strlen} \i{String Handling in Macros}: \i\c{%strlen} and \i\c{%substr}
2014 It's often useful to be able to handle strings in macros. NASM
2015 supports two simple string handling macro operators from which
2016 more complex operations can be constructed.
2019 \S{strlen} \i{String Length}: \i\c{%strlen}
2021 The \c{%strlen} macro is like \c{%assign} macro in that it creates
2022 (or redefines) a numeric value to a macro. The difference is that
2023 with \c{%strlen}, the numeric value is the length of a string. An
2024 example of the use of this would be:
2026 \c %strlen charcnt 'my string'
2028 In this example, \c{charcnt} would receive the value 8, just as
2029 if an \c{%assign} had been used. In this example, \c{'my string'}
2030 was a literal string but it could also have been a single-line
2031 macro that expands to a string, as in the following example:
2033 \c %define sometext 'my string'
2034 \c %strlen charcnt sometext
2036 As in the first case, this would result in \c{charcnt} being
2037 assigned the value of 8.
2040 \S{substr} \i{Sub-strings}: \i\c{%substr}
2042 Individual letters in strings can be extracted using \c{%substr}.
2043 An example of its use is probably more useful than the description:
2045 \c %substr mychar 'xyz' 1 ; equivalent to %define mychar 'x'
2046 \c %substr mychar 'xyz' 2 ; equivalent to %define mychar 'y'
2047 \c %substr mychar 'xyz' 3 ; equivalent to %define mychar 'z'
2049 In this example, mychar gets the value of 'y'. As with \c{%strlen}
2050 (see \k{strlen}), the first parameter is the single-line macro to
2051 be created and the second is the string. The third parameter
2052 specifies which character is to be selected. Note that the first
2053 index is 1, not 0 and the last index is equal to the value that
2054 \c{%strlen} would assign given the same string. Index values out
2055 of range result in an empty string.
2058 \H{mlmacro} \i{Multi-Line Macros}: \I\c{%imacro}\i\c{%macro}
2060 Multi-line macros are much more like the type of macro seen in MASM
2061 and TASM: a multi-line macro definition in NASM looks something like
2064 \c %macro prologue 1
2072 This defines a C-like function prologue as a macro: so you would
2073 invoke the macro with a call such as
2075 \c myfunc: prologue 12
2077 which would expand to the three lines of code
2083 The number \c{1} after the macro name in the \c{%macro} line defines
2084 the number of parameters the macro \c{prologue} expects to receive.
2085 The use of \c{%1} inside the macro definition refers to the first
2086 parameter to the macro call. With a macro taking more than one
2087 parameter, subsequent parameters would be referred to as \c{%2},
2090 Multi-line macros, like single-line macros, are \i{case-sensitive},
2091 unless you define them using the alternative directive \c{%imacro}.
2093 If you need to pass a comma as \e{part} of a parameter to a
2094 multi-line macro, you can do that by enclosing the entire parameter
2095 in \I{braces, around macro parameters}braces. So you could code
2104 \c silly 'a', letter_a ; letter_a: db 'a'
2105 \c silly 'ab', string_ab ; string_ab: db 'ab'
2106 \c silly {13,10}, crlf ; crlf: db 13,10
2109 \S{mlmacover} Overloading Multi-Line Macros\I{overloading, multi-line macros}
2111 As with single-line macros, multi-line macros can be overloaded by
2112 defining the same macro name several times with different numbers of
2113 parameters. This time, no exception is made for macros with no
2114 parameters at all. So you could define
2116 \c %macro prologue 0
2123 to define an alternative form of the function prologue which
2124 allocates no local stack space.
2126 Sometimes, however, you might want to `overload' a machine
2127 instruction; for example, you might want to define
2136 so that you could code
2138 \c push ebx ; this line is not a macro call
2139 \c push eax,ecx ; but this one is
2141 Ordinarily, NASM will give a warning for the first of the above two
2142 lines, since \c{push} is now defined to be a macro, and is being
2143 invoked with a number of parameters for which no definition has been
2144 given. The correct code will still be generated, but the assembler
2145 will give a warning. This warning can be disabled by the use of the
2146 \c{-w-macro-params} command-line option (see \k{opt-w}).
2149 \S{maclocal} \i{Macro-Local Labels}
2151 NASM allows you to define labels within a multi-line macro
2152 definition in such a way as to make them local to the macro call: so
2153 calling the same macro multiple times will use a different label
2154 each time. You do this by prefixing \i\c{%%} to the label name. So
2155 you can invent an instruction which executes a \c{RET} if the \c{Z}
2156 flag is set by doing this:
2166 You can call this macro as many times as you want, and every time
2167 you call it NASM will make up a different `real' name to substitute
2168 for the label \c{%%skip}. The names NASM invents are of the form
2169 \c{..@2345.skip}, where the number 2345 changes with every macro
2170 call. The \i\c{..@} prefix prevents macro-local labels from
2171 interfering with the local label mechanism, as described in
2172 \k{locallab}. You should avoid defining your own labels in this form
2173 (the \c{..@} prefix, then a number, then another period) in case
2174 they interfere with macro-local labels.
2177 \S{mlmacgre} \i{Greedy Macro Parameters}
2179 Occasionally it is useful to define a macro which lumps its entire
2180 command line into one parameter definition, possibly after
2181 extracting one or two smaller parameters from the front. An example
2182 might be a macro to write a text string to a file in MS-DOS, where
2183 you might want to be able to write
2185 \c writefile [filehandle],"hello, world",13,10
2187 NASM allows you to define the last parameter of a macro to be
2188 \e{greedy}, meaning that if you invoke the macro with more
2189 parameters than it expects, all the spare parameters get lumped into
2190 the last defined one along with the separating commas. So if you
2193 \c %macro writefile 2+
2199 \c mov cx,%%endstr-%%str
2206 then the example call to \c{writefile} above will work as expected:
2207 the text before the first comma, \c{[filehandle]}, is used as the
2208 first macro parameter and expanded when \c{%1} is referred to, and
2209 all the subsequent text is lumped into \c{%2} and placed after the
2212 The greedy nature of the macro is indicated to NASM by the use of
2213 the \I{+ modifier}\c{+} sign after the parameter count on the
2216 If you define a greedy macro, you are effectively telling NASM how
2217 it should expand the macro given \e{any} number of parameters from
2218 the actual number specified up to infinity; in this case, for
2219 example, NASM now knows what to do when it sees a call to
2220 \c{writefile} with 2, 3, 4 or more parameters. NASM will take this
2221 into account when overloading macros, and will not allow you to
2222 define another form of \c{writefile} taking 4 parameters (for
2225 Of course, the above macro could have been implemented as a
2226 non-greedy macro, in which case the call to it would have had to
2229 \c writefile [filehandle], {"hello, world",13,10}
2231 NASM provides both mechanisms for putting \i{commas in macro
2232 parameters}, and you choose which one you prefer for each macro
2235 See \k{sectmac} for a better way to write the above macro.
2238 \S{mlmacdef} \i{Default Macro Parameters}
2240 NASM also allows you to define a multi-line macro with a \e{range}
2241 of allowable parameter counts. If you do this, you can specify
2242 defaults for \i{omitted parameters}. So, for example:
2244 \c %macro die 0-1 "Painful program death has occurred."
2252 This macro (which makes use of the \c{writefile} macro defined in
2253 \k{mlmacgre}) can be called with an explicit error message, which it
2254 will display on the error output stream before exiting, or it can be
2255 called with no parameters, in which case it will use the default
2256 error message supplied in the macro definition.
2258 In general, you supply a minimum and maximum number of parameters
2259 for a macro of this type; the minimum number of parameters are then
2260 required in the macro call, and then you provide defaults for the
2261 optional ones. So if a macro definition began with the line
2263 \c %macro foobar 1-3 eax,[ebx+2]
2265 then it could be called with between one and three parameters, and
2266 \c{%1} would always be taken from the macro call. \c{%2}, if not
2267 specified by the macro call, would default to \c{eax}, and \c{%3} if
2268 not specified would default to \c{[ebx+2]}.
2270 You may omit parameter defaults from the macro definition, in which
2271 case the parameter default is taken to be blank. This can be useful
2272 for macros which can take a variable number of parameters, since the
2273 \i\c{%0} token (see \k{percent0}) allows you to determine how many
2274 parameters were really passed to the macro call.
2276 This defaulting mechanism can be combined with the greedy-parameter
2277 mechanism; so the \c{die} macro above could be made more powerful,
2278 and more useful, by changing the first line of the definition to
2280 \c %macro die 0-1+ "Painful program death has occurred.",13,10
2282 The maximum parameter count can be infinite, denoted by \c{*}. In
2283 this case, of course, it is impossible to provide a \e{full} set of
2284 default parameters. Examples of this usage are shown in \k{rotate}.
2287 \S{percent0} \i\c{%0}: \I{counting macro parameters}Macro Parameter Counter
2289 For a macro which can take a variable number of parameters, the
2290 parameter reference \c{%0} will return a numeric constant giving the
2291 number of parameters passed to the macro. This can be used as an
2292 argument to \c{%rep} (see \k{rep}) in order to iterate through all
2293 the parameters of a macro. Examples are given in \k{rotate}.
2296 \S{rotate} \i\c{%rotate}: \i{Rotating Macro Parameters}
2298 Unix shell programmers will be familiar with the \I{shift
2299 command}\c{shift} shell command, which allows the arguments passed
2300 to a shell script (referenced as \c{$1}, \c{$2} and so on) to be
2301 moved left by one place, so that the argument previously referenced
2302 as \c{$2} becomes available as \c{$1}, and the argument previously
2303 referenced as \c{$1} is no longer available at all.
2305 NASM provides a similar mechanism, in the form of \c{%rotate}. As
2306 its name suggests, it differs from the Unix \c{shift} in that no
2307 parameters are lost: parameters rotated off the left end of the
2308 argument list reappear on the right, and vice versa.
2310 \c{%rotate} is invoked with a single numeric argument (which may be
2311 an expression). The macro parameters are rotated to the left by that
2312 many places. If the argument to \c{%rotate} is negative, the macro
2313 parameters are rotated to the right.
2315 \I{iterating over macro parameters}So a pair of macros to save and
2316 restore a set of registers might work as follows:
2318 \c %macro multipush 1-*
2327 This macro invokes the \c{PUSH} instruction on each of its arguments
2328 in turn, from left to right. It begins by pushing its first
2329 argument, \c{%1}, then invokes \c{%rotate} to move all the arguments
2330 one place to the left, so that the original second argument is now
2331 available as \c{%1}. Repeating this procedure as many times as there
2332 were arguments (achieved by supplying \c{%0} as the argument to
2333 \c{%rep}) causes each argument in turn to be pushed.
2335 Note also the use of \c{*} as the maximum parameter count,
2336 indicating that there is no upper limit on the number of parameters
2337 you may supply to the \i\c{multipush} macro.
2339 It would be convenient, when using this macro, to have a \c{POP}
2340 equivalent, which \e{didn't} require the arguments to be given in
2341 reverse order. Ideally, you would write the \c{multipush} macro
2342 call, then cut-and-paste the line to where the pop needed to be
2343 done, and change the name of the called macro to \c{multipop}, and
2344 the macro would take care of popping the registers in the opposite
2345 order from the one in which they were pushed.
2347 This can be done by the following definition:
2349 \c %macro multipop 1-*
2358 This macro begins by rotating its arguments one place to the
2359 \e{right}, so that the original \e{last} argument appears as \c{%1}.
2360 This is then popped, and the arguments are rotated right again, so
2361 the second-to-last argument becomes \c{%1}. Thus the arguments are
2362 iterated through in reverse order.
2365 \S{concat} \i{Concatenating Macro Parameters}
2367 NASM can concatenate macro parameters on to other text surrounding
2368 them. This allows you to declare a family of symbols, for example,
2369 in a macro definition. If, for example, you wanted to generate a
2370 table of key codes along with offsets into the table, you could code
2373 \c %macro keytab_entry 2
2375 \c keypos%1 equ $-keytab
2381 \c keytab_entry F1,128+1
2382 \c keytab_entry F2,128+2
2383 \c keytab_entry Return,13
2385 which would expand to
2388 \c keyposF1 equ $-keytab
2390 \c keyposF2 equ $-keytab
2392 \c keyposReturn equ $-keytab
2395 You can just as easily concatenate text on to the other end of a
2396 macro parameter, by writing \c{%1foo}.
2398 If you need to append a \e{digit} to a macro parameter, for example
2399 defining labels \c{foo1} and \c{foo2} when passed the parameter
2400 \c{foo}, you can't code \c{%11} because that would be taken as the
2401 eleventh macro parameter. Instead, you must code
2402 \I{braces, after % sign}\c{%\{1\}1}, which will separate the first
2403 \c{1} (giving the number of the macro parameter) from the second
2404 (literal text to be concatenated to the parameter).
2406 This concatenation can also be applied to other preprocessor in-line
2407 objects, such as macro-local labels (\k{maclocal}) and context-local
2408 labels (\k{ctxlocal}). In all cases, ambiguities in syntax can be
2409 resolved by enclosing everything after the \c{%} sign and before the
2410 literal text in braces: so \c{%\{%foo\}bar} concatenates the text
2411 \c{bar} to the end of the real name of the macro-local label
2412 \c{%%foo}. (This is unnecessary, since the form NASM uses for the
2413 real names of macro-local labels means that the two usages
2414 \c{%\{%foo\}bar} and \c{%%foobar} would both expand to the same
2415 thing anyway; nevertheless, the capability is there.)
2418 \S{mlmaccc} \i{Condition Codes as Macro Parameters}
2420 NASM can give special treatment to a macro parameter which contains
2421 a condition code. For a start, you can refer to the macro parameter
2422 \c{%1} by means of the alternative syntax \i\c{%+1}, which informs
2423 NASM that this macro parameter is supposed to contain a condition
2424 code, and will cause the preprocessor to report an error message if
2425 the macro is called with a parameter which is \e{not} a valid
2428 Far more usefully, though, you can refer to the macro parameter by
2429 means of \i\c{%-1}, which NASM will expand as the \e{inverse}
2430 condition code. So the \c{retz} macro defined in \k{maclocal} can be
2431 replaced by a general \i{conditional-return macro} like this:
2441 This macro can now be invoked using calls like \c{retc ne}, which
2442 will cause the conditional-jump instruction in the macro expansion
2443 to come out as \c{JE}, or \c{retc po} which will make the jump a
2446 The \c{%+1} macro-parameter reference is quite happy to interpret
2447 the arguments \c{CXZ} and \c{ECXZ} as valid condition codes;
2448 however, \c{%-1} will report an error if passed either of these,
2449 because no inverse condition code exists.
2452 \S{nolist} \i{Disabling Listing Expansion}\I\c{.nolist}
2454 When NASM is generating a listing file from your program, it will
2455 generally expand multi-line macros by means of writing the macro
2456 call and then listing each line of the expansion. This allows you to
2457 see which instructions in the macro expansion are generating what
2458 code; however, for some macros this clutters the listing up
2461 NASM therefore provides the \c{.nolist} qualifier, which you can
2462 include in a macro definition to inhibit the expansion of the macro
2463 in the listing file. The \c{.nolist} qualifier comes directly after
2464 the number of parameters, like this:
2466 \c %macro foo 1.nolist
2470 \c %macro bar 1-5+.nolist a,b,c,d,e,f,g,h
2472 \H{condasm} \i{Conditional Assembly}\I\c{%if}
2474 Similarly to the C preprocessor, NASM allows sections of a source
2475 file to be assembled only if certain conditions are met. The general
2476 syntax of this feature looks like this:
2479 \c ; some code which only appears if <condition> is met
2480 \c %elif<condition2>
2481 \c ; only appears if <condition> is not met but <condition2> is
2483 \c ; this appears if neither <condition> nor <condition2> was met
2486 The \i\c{%else} clause is optional, as is the \i\c{%elif} clause.
2487 You can have more than one \c{%elif} clause as well.
2490 \S{ifdef} \i\c{%ifdef}: Testing Single-Line Macro Existence\I{testing,
2491 single-line macro existence}
2493 Beginning a conditional-assembly block with the line \c{%ifdef
2494 MACRO} will assemble the subsequent code if, and only if, a
2495 single-line macro called \c{MACRO} is defined. If not, then the
2496 \c{%elif} and \c{%else} blocks (if any) will be processed instead.
2498 For example, when debugging a program, you might want to write code
2501 \c ; perform some function
2503 \c writefile 2,"Function performed successfully",13,10
2505 \c ; go and do something else
2507 Then you could use the command-line option \c{-dDEBUG} to create a
2508 version of the program which produced debugging messages, and remove
2509 the option to generate the final release version of the program.
2511 You can test for a macro \e{not} being defined by using
2512 \i\c{%ifndef} instead of \c{%ifdef}. You can also test for macro
2513 definitions in \c{%elif} blocks by using \i\c{%elifdef} and
2517 \S{ifmacro} \i\c{ifmacro}: Testing Multi-Line Macro
2518 Existence\I{testing, multi-line macro existence}
2520 The \c{%ifmacro} directive operates in the same way as the \c{%ifdef}
2521 directive, except that it checks for the existence of a multi-line macro.
2523 For example, you may be working with a large project and not have control
2524 over the macros in a library. You may want to create a macro with one
2525 name if it doesn't already exist, and another name if one with that name
2528 The \c{%ifmacro} is considered true if defining a macro with the given name
2529 and number of arguments would cause a definitions conflict. For example:
2531 \c %ifmacro MyMacro 1-3
2533 \c %error "MyMacro 1-3" causes a conflict with an existing macro.
2537 \c %macro MyMacro 1-3
2539 \c ; insert code to define the macro
2545 This will create the macro "MyMacro 1-3" if no macro already exists which
2546 would conflict with it, and emits a warning if there would be a definition
2549 You can test for the macro not existing by using the \i\c{%ifnmacro} instead
2550 of \c{%ifmacro}. Additional tests can be performed in \c{%elif} blocks by using
2551 \i\c{%elifmacro} and \i\c{%elifnmacro}.
2554 \S{ifctx} \i\c{%ifctx}: Testing the Context Stack\I{testing, context
2557 The conditional-assembly construct \c{%ifctx ctxname} will cause the
2558 subsequent code to be assembled if and only if the top context on
2559 the preprocessor's context stack has the name \c{ctxname}. As with
2560 \c{%ifdef}, the inverse and \c{%elif} forms \i\c{%ifnctx},
2561 \i\c{%elifctx} and \i\c{%elifnctx} are also supported.
2563 For more details of the context stack, see \k{ctxstack}. For a
2564 sample use of \c{%ifctx}, see \k{blockif}.
2567 \S{if} \i\c{%if}: Testing Arbitrary Numeric Expressions\I{testing,
2568 arbitrary numeric expressions}
2570 The conditional-assembly construct \c{%if expr} will cause the
2571 subsequent code to be assembled if and only if the value of the
2572 numeric expression \c{expr} is non-zero. An example of the use of
2573 this feature is in deciding when to break out of a \c{%rep}
2574 preprocessor loop: see \k{rep} for a detailed example.
2576 The expression given to \c{%if}, and its counterpart \i\c{%elif}, is
2577 a critical expression (see \k{crit}).
2579 \c{%if} extends the normal NASM expression syntax, by providing a
2580 set of \i{relational operators} which are not normally available in
2581 expressions. The operators \i\c{=}, \i\c{<}, \i\c{>}, \i\c{<=},
2582 \i\c{>=} and \i\c{<>} test equality, less-than, greater-than,
2583 less-or-equal, greater-or-equal and not-equal respectively. The
2584 C-like forms \i\c{==} and \i\c{!=} are supported as alternative
2585 forms of \c{=} and \c{<>}. In addition, low-priority logical
2586 operators \i\c{&&}, \i\c{^^} and \i\c{||} are provided, supplying
2587 \i{logical AND}, \i{logical XOR} and \i{logical OR}. These work like
2588 the C logical operators (although C has no logical XOR), in that
2589 they always return either 0 or 1, and treat any non-zero input as 1
2590 (so that \c{^^}, for example, returns 1 if exactly one of its inputs
2591 is zero, and 0 otherwise). The relational operators also return 1
2592 for true and 0 for false.
2594 Like most other \c{%if} constructs, \c{%if} has a counterpart
2595 \i\c{%elif}, and negative forms \i\c{%ifn} and \i\c{%elifn}.
2597 \S{ifidn} \i\c{%ifidn} and \i\c{%ifidni}: Testing Exact Text
2598 Identity\I{testing, exact text identity}
2600 The construct \c{%ifidn text1,text2} will cause the subsequent code
2601 to be assembled if and only if \c{text1} and \c{text2}, after
2602 expanding single-line macros, are identical pieces of text.
2603 Differences in white space are not counted.
2605 \c{%ifidni} is similar to \c{%ifidn}, but is \i{case-insensitive}.
2607 For example, the following macro pushes a register or number on the
2608 stack, and allows you to treat \c{IP} as a real register:
2610 \c %macro pushparam 1
2621 Like most other \c{%if} constructs, \c{%ifidn} has a counterpart
2622 \i\c{%elifidn}, and negative forms \i\c{%ifnidn} and \i\c{%elifnidn}.
2623 Similarly, \c{%ifidni} has counterparts \i\c{%elifidni},
2624 \i\c{%ifnidni} and \i\c{%elifnidni}.
2627 \S{iftyp} \i\c{%ifid}, \i\c{%ifnum}, \i\c{%ifstr}: Testing Token
2628 Types\I{testing, token types}
2630 Some macros will want to perform different tasks depending on
2631 whether they are passed a number, a string, or an identifier. For
2632 example, a string output macro might want to be able to cope with
2633 being passed either a string constant or a pointer to an existing
2636 The conditional assembly construct \c{%ifid}, taking one parameter
2637 (which may be blank), assembles the subsequent code if and only if
2638 the first token in the parameter exists and is an identifier.
2639 \c{%ifnum} works similarly, but tests for the token being a numeric
2640 constant; \c{%ifstr} tests for it being a string.
2642 For example, the \c{writefile} macro defined in \k{mlmacgre} can be
2643 extended to take advantage of \c{%ifstr} in the following fashion:
2645 \c %macro writefile 2-3+
2654 \c %%endstr: mov dx,%%str
2655 \c mov cx,%%endstr-%%str
2666 Then the \c{writefile} macro can cope with being called in either of
2667 the following two ways:
2669 \c writefile [file], strpointer, length
2670 \c writefile [file], "hello", 13, 10
2672 In the first, \c{strpointer} is used as the address of an
2673 already-declared string, and \c{length} is used as its length; in
2674 the second, a string is given to the macro, which therefore declares
2675 it itself and works out the address and length for itself.
2677 Note the use of \c{%if} inside the \c{%ifstr}: this is to detect
2678 whether the macro was passed two arguments (so the string would be a
2679 single string constant, and \c{db %2} would be adequate) or more (in
2680 which case, all but the first two would be lumped together into
2681 \c{%3}, and \c{db %2,%3} would be required).
2683 \I\c{%ifnid}\I\c{%elifid}\I\c{%elifnid}\I\c{%ifnnum}\I\c{%elifnum}
2684 \I\c{%elifnnum}\I\c{%ifnstr}\I\c{%elifstr}\I\c{%elifnstr}
2685 The usual \c{%elifXXX}, \c{%ifnXXX} and \c{%elifnXXX} versions exist
2686 for each of \c{%ifid}, \c{%ifnum} and \c{%ifstr}.
2689 \S{pperror} \i\c{%error}: Reporting \i{User-Defined Errors}
2691 The preprocessor directive \c{%error} will cause NASM to report an
2692 error if it occurs in assembled code. So if other users are going to
2693 try to assemble your source files, you can ensure that they define
2694 the right macros by means of code like this:
2696 \c %ifdef SOME_MACRO
2698 \c %elifdef SOME_OTHER_MACRO
2699 \c ; do some different setup
2701 \c %error Neither SOME_MACRO nor SOME_OTHER_MACRO was defined.
2704 Then any user who fails to understand the way your code is supposed
2705 to be assembled will be quickly warned of their mistake, rather than
2706 having to wait until the program crashes on being run and then not
2707 knowing what went wrong.
2710 \H{rep} \i{Preprocessor Loops}\I{repeating code}: \i\c{%rep}
2712 NASM's \c{TIMES} prefix, though useful, cannot be used to invoke a
2713 multi-line macro multiple times, because it is processed by NASM
2714 after macros have already been expanded. Therefore NASM provides
2715 another form of loop, this time at the preprocessor level: \c{%rep}.
2717 The directives \c{%rep} and \i\c{%endrep} (\c{%rep} takes a numeric
2718 argument, which can be an expression; \c{%endrep} takes no
2719 arguments) can be used to enclose a chunk of code, which is then
2720 replicated as many times as specified by the preprocessor:
2724 \c inc word [table+2*i]
2728 This will generate a sequence of 64 \c{INC} instructions,
2729 incrementing every word of memory from \c{[table]} to
2732 For more complex termination conditions, or to break out of a repeat
2733 loop part way along, you can use the \i\c{%exitrep} directive to
2734 terminate the loop, like this:
2749 \c fib_number equ ($-fibonacci)/2
2751 This produces a list of all the Fibonacci numbers that will fit in
2752 16 bits. Note that a maximum repeat count must still be given to
2753 \c{%rep}. This is to prevent the possibility of NASM getting into an
2754 infinite loop in the preprocessor, which (on multitasking or
2755 multi-user systems) would typically cause all the system memory to
2756 be gradually used up and other applications to start crashing.
2759 \H{include} \i{Including Other Files}
2761 Using, once again, a very similar syntax to the C preprocessor,
2762 NASM's preprocessor lets you include other source files into your
2763 code. This is done by the use of the \i\c{%include} directive:
2765 \c %include "macros.mac"
2767 will include the contents of the file \c{macros.mac} into the source
2768 file containing the \c{%include} directive.
2770 Include files are \I{searching for include files}searched for in the
2771 current directory (the directory you're in when you run NASM, as
2772 opposed to the location of the NASM executable or the location of
2773 the source file), plus any directories specified on the NASM command
2774 line using the \c{-i} option.
2776 The standard C idiom for preventing a file being included more than
2777 once is just as applicable in NASM: if the file \c{macros.mac} has
2780 \c %ifndef MACROS_MAC
2781 \c %define MACROS_MAC
2782 \c ; now define some macros
2785 then including the file more than once will not cause errors,
2786 because the second time the file is included nothing will happen
2787 because the macro \c{MACROS_MAC} will already be defined.
2789 You can force a file to be included even if there is no \c{%include}
2790 directive that explicitly includes it, by using the \i\c{-p} option
2791 on the NASM command line (see \k{opt-p}).
2794 \H{ctxstack} The \i{Context Stack}
2796 Having labels that are local to a macro definition is sometimes not
2797 quite powerful enough: sometimes you want to be able to share labels
2798 between several macro calls. An example might be a \c{REPEAT} ...
2799 \c{UNTIL} loop, in which the expansion of the \c{REPEAT} macro
2800 would need to be able to refer to a label which the \c{UNTIL} macro
2801 had defined. However, for such a macro you would also want to be
2802 able to nest these loops.
2804 NASM provides this level of power by means of a \e{context stack}.
2805 The preprocessor maintains a stack of \e{contexts}, each of which is
2806 characterized by a name. You add a new context to the stack using
2807 the \i\c{%push} directive, and remove one using \i\c{%pop}. You can
2808 define labels that are local to a particular context on the stack.
2811 \S{pushpop} \i\c{%push} and \i\c{%pop}: \I{creating
2812 contexts}\I{removing contexts}Creating and Removing Contexts
2814 The \c{%push} directive is used to create a new context and place it
2815 on the top of the context stack. \c{%push} requires one argument,
2816 which is the name of the context. For example:
2820 This pushes a new context called \c{foobar} on the stack. You can
2821 have several contexts on the stack with the same name: they can
2822 still be distinguished.
2824 The directive \c{%pop}, requiring no arguments, removes the top
2825 context from the context stack and destroys it, along with any
2826 labels associated with it.
2829 \S{ctxlocal} \i{Context-Local Labels}
2831 Just as the usage \c{%%foo} defines a label which is local to the
2832 particular macro call in which it is used, the usage \I{%$}\c{%$foo}
2833 is used to define a label which is local to the context on the top
2834 of the context stack. So the \c{REPEAT} and \c{UNTIL} example given
2835 above could be implemented by means of:
2851 and invoked by means of, for example,
2859 which would scan every fourth byte of a string in search of the byte
2862 If you need to define, or access, labels local to the context
2863 \e{below} the top one on the stack, you can use \I{%$$}\c{%$$foo}, or
2864 \c{%$$$foo} for the context below that, and so on.
2867 \S{ctxdefine} \i{Context-Local Single-Line Macros}
2869 NASM also allows you to define single-line macros which are local to
2870 a particular context, in just the same way:
2872 \c %define %$localmac 3
2874 will define the single-line macro \c{%$localmac} to be local to the
2875 top context on the stack. Of course, after a subsequent \c{%push},
2876 it can then still be accessed by the name \c{%$$localmac}.
2879 \S{ctxrepl} \i\c{%repl}: \I{renaming contexts}Renaming a Context
2881 If you need to change the name of the top context on the stack (in
2882 order, for example, to have it respond differently to \c{%ifctx}),
2883 you can execute a \c{%pop} followed by a \c{%push}; but this will
2884 have the side effect of destroying all context-local labels and
2885 macros associated with the context that was just popped.
2887 NASM provides the directive \c{%repl}, which \e{replaces} a context
2888 with a different name, without touching the associated macros and
2889 labels. So you could replace the destructive code
2894 with the non-destructive version \c{%repl newname}.
2897 \S{blockif} Example Use of the \i{Context Stack}: \i{Block IFs}
2899 This example makes use of almost all the context-stack features,
2900 including the conditional-assembly construct \i\c{%ifctx}, to
2901 implement a block IF statement as a set of macros.
2917 \c %error "expected `if' before `else'"
2931 \c %error "expected `if' or `else' before `endif'"
2936 This code is more robust than the \c{REPEAT} and \c{UNTIL} macros
2937 given in \k{ctxlocal}, because it uses conditional assembly to check
2938 that the macros are issued in the right order (for example, not
2939 calling \c{endif} before \c{if}) and issues a \c{%error} if they're
2942 In addition, the \c{endif} macro has to be able to cope with the two
2943 distinct cases of either directly following an \c{if}, or following
2944 an \c{else}. It achieves this, again, by using conditional assembly
2945 to do different things depending on whether the context on top of
2946 the stack is \c{if} or \c{else}.
2948 The \c{else} macro has to preserve the context on the stack, in
2949 order to have the \c{%$ifnot} referred to by the \c{if} macro be the
2950 same as the one defined by the \c{endif} macro, but has to change
2951 the context's name so that \c{endif} will know there was an
2952 intervening \c{else}. It does this by the use of \c{%repl}.
2954 A sample usage of these macros might look like:
2976 The block-\c{IF} macros handle nesting quite happily, by means of
2977 pushing another context, describing the inner \c{if}, on top of the
2978 one describing the outer \c{if}; thus \c{else} and \c{endif} always
2979 refer to the last unmatched \c{if} or \c{else}.
2982 \H{stdmac} \i{Standard Macros}
2984 NASM defines a set of standard macros, which are already defined
2985 when it starts to process any source file. If you really need a
2986 program to be assembled with no pre-defined macros, you can use the
2987 \i\c{%clear} directive to empty the preprocessor of everything but
2988 context-local preprocessor variables and single-line macros.
2990 Most \i{user-level assembler directives} (see \k{directive}) are
2991 implemented as macros which invoke primitive directives; these are
2992 described in \k{directive}. The rest of the standard macro set is
2996 \S{stdmacver} \i\c{__NASM_MAJOR__}, \i\c{__NASM_MINOR__},
2997 \i\c{__NASM_SUBMINOR__} and \i\c{___NASM_PATCHLEVEL__}: \i{NASM Version}
2999 The single-line macros \c{__NASM_MAJOR__}, \c{__NASM_MINOR__},
3000 \c{__NASM_SUBMINOR__} and \c{___NASM_PATCHLEVEL__} expand to the
3001 major, minor, subminor and patch level parts of the \i{version
3002 number of NASM} being used. So, under NASM 0.98.32p1 for
3003 example, \c{__NASM_MAJOR__} would be defined to be 0, \c{__NASM_MINOR__}
3004 would be defined as 98, \c{__NASM_SUBMINOR__} would be defined to 32,
3005 and \c{___NASM_PATCHLEVEL__} would be defined as 1.
3008 \S{stdmacverid} \i\c{__NASM_VERSION_ID__}: \i{NASM Version ID}
3010 The single-line macro \c{__NASM_VERSION_ID__} expands to a dword integer
3011 representing the full version number of the version of nasm being used.
3012 The value is the equivalent to \c{__NASM_MAJOR__}, \c{__NASM_MINOR__},
3013 \c{__NASM_SUBMINOR__} and \c{___NASM_PATCHLEVEL__} concatenated to
3014 produce a single doubleword. Hence, for 0.98.32p1, the returned number
3015 would be equivalent to:
3023 Note that the above lines are generate exactly the same code, the second
3024 line is used just to give an indication of the order that the separate
3025 values will be present in memory.
3028 \S{stdmacverstr} \i\c{__NASM_VER__}: \i{NASM Version string}
3030 The single-line macro \c{__NASM_VER__} expands to a string which defines
3031 the version number of nasm being used. So, under NASM 0.98.32 for example,
3040 \S{fileline} \i\c{__FILE__} and \i\c{__LINE__}: File Name and Line Number
3042 Like the C preprocessor, NASM allows the user to find out the file
3043 name and line number containing the current instruction. The macro
3044 \c{__FILE__} expands to a string constant giving the name of the
3045 current input file (which may change through the course of assembly
3046 if \c{%include} directives are used), and \c{__LINE__} expands to a
3047 numeric constant giving the current line number in the input file.
3049 These macros could be used, for example, to communicate debugging
3050 information to a macro, since invoking \c{__LINE__} inside a macro
3051 definition (either single-line or multi-line) will return the line
3052 number of the macro \e{call}, rather than \e{definition}. So to
3053 determine where in a piece of code a crash is occurring, for
3054 example, one could write a routine \c{stillhere}, which is passed a
3055 line number in \c{EAX} and outputs something like `line 155: still
3056 here'. You could then write a macro
3058 \c %macro notdeadyet 0
3067 and then pepper your code with calls to \c{notdeadyet} until you
3068 find the crash point.
3070 \S{bitsm} \i\c{__BITS__}: Current BITS Mode
3072 The \c{__BITS__} standard macro is updated every time that the BITS mode is
3073 set using the \c{BITS XX} or \c{[BITS XX]} directive, where XX is a valid mode
3074 number of 16, 32 or 64. \c{__BITS__} receives the specified mode number and
3075 makes it globally available. This can be very useful for those who utilize
3076 mode-dependent macros.
3079 \S{struc} \i\c{STRUC} and \i\c{ENDSTRUC}: \i{Declaring Structure} Data Types
3081 The core of NASM contains no intrinsic means of defining data
3082 structures; instead, the preprocessor is sufficiently powerful that
3083 data structures can be implemented as a set of macros. The macros
3084 \c{STRUC} and \c{ENDSTRUC} are used to define a structure data type.
3086 \c{STRUC} takes one parameter, which is the name of the data type.
3087 This name is defined as a symbol with the value zero, and also has
3088 the suffix \c{_size} appended to it and is then defined as an
3089 \c{EQU} giving the size of the structure. Once \c{STRUC} has been
3090 issued, you are defining the structure, and should define fields
3091 using the \c{RESB} family of pseudo-instructions, and then invoke
3092 \c{ENDSTRUC} to finish the definition.
3094 For example, to define a structure called \c{mytype} containing a
3095 longword, a word, a byte and a string of bytes, you might code
3106 The above code defines six symbols: \c{mt_long} as 0 (the offset
3107 from the beginning of a \c{mytype} structure to the longword field),
3108 \c{mt_word} as 4, \c{mt_byte} as 6, \c{mt_str} as 7, \c{mytype_size}
3109 as 39, and \c{mytype} itself as zero.
3111 The reason why the structure type name is defined at zero is a side
3112 effect of allowing structures to work with the local label
3113 mechanism: if your structure members tend to have the same names in
3114 more than one structure, you can define the above structure like this:
3125 This defines the offsets to the structure fields as \c{mytype.long},
3126 \c{mytype.word}, \c{mytype.byte} and \c{mytype.str}.
3128 NASM, since it has no \e{intrinsic} structure support, does not
3129 support any form of period notation to refer to the elements of a
3130 structure once you have one (except the above local-label notation),
3131 so code such as \c{mov ax,[mystruc.mt_word]} is not valid.
3132 \c{mt_word} is a constant just like any other constant, so the
3133 correct syntax is \c{mov ax,[mystruc+mt_word]} or \c{mov
3134 ax,[mystruc+mytype.word]}.
3137 \S{istruc} \i\c{ISTRUC}, \i\c{AT} and \i\c{IEND}: Declaring
3138 \i{Instances of Structures}
3140 Having defined a structure type, the next thing you typically want
3141 to do is to declare instances of that structure in your data
3142 segment. NASM provides an easy way to do this in the \c{ISTRUC}
3143 mechanism. To declare a structure of type \c{mytype} in a program,
3144 you code something like this:
3149 \c at mt_long, dd 123456
3150 \c at mt_word, dw 1024
3151 \c at mt_byte, db 'x'
3152 \c at mt_str, db 'hello, world', 13, 10, 0
3156 The function of the \c{AT} macro is to make use of the \c{TIMES}
3157 prefix to advance the assembly position to the correct point for the
3158 specified structure field, and then to declare the specified data.
3159 Therefore the structure fields must be declared in the same order as
3160 they were specified in the structure definition.
3162 If the data to go in a structure field requires more than one source
3163 line to specify, the remaining source lines can easily come after
3164 the \c{AT} line. For example:
3166 \c at mt_str, db 123,134,145,156,167,178,189
3169 Depending on personal taste, you can also omit the code part of the
3170 \c{AT} line completely, and start the structure field on the next
3174 \c db 'hello, world'
3178 \S{align} \i\c{ALIGN} and \i\c{ALIGNB}: Data Alignment
3180 The \c{ALIGN} and \c{ALIGNB} macros provides a convenient way to
3181 align code or data on a word, longword, paragraph or other boundary.
3182 (Some assemblers call this directive \i\c{EVEN}.) The syntax of the
3183 \c{ALIGN} and \c{ALIGNB} macros is
3185 \c align 4 ; align on 4-byte boundary
3186 \c align 16 ; align on 16-byte boundary
3187 \c align 8,db 0 ; pad with 0s rather than NOPs
3188 \c align 4,resb 1 ; align to 4 in the BSS
3189 \c alignb 4 ; equivalent to previous line
3191 Both macros require their first argument to be a power of two; they
3192 both compute the number of additional bytes required to bring the
3193 length of the current section up to a multiple of that power of two,
3194 and then apply the \c{TIMES} prefix to their second argument to
3195 perform the alignment.
3197 If the second argument is not specified, the default for \c{ALIGN}
3198 is \c{NOP}, and the default for \c{ALIGNB} is \c{RESB 1}. So if the
3199 second argument is specified, the two macros are equivalent.
3200 Normally, you can just use \c{ALIGN} in code and data sections and
3201 \c{ALIGNB} in BSS sections, and never need the second argument
3202 except for special purposes.
3204 \c{ALIGN} and \c{ALIGNB}, being simple macros, perform no error
3205 checking: they cannot warn you if their first argument fails to be a
3206 power of two, or if their second argument generates more than one
3207 byte of code. In each of these cases they will silently do the wrong
3210 \c{ALIGNB} (or \c{ALIGN} with a second argument of \c{RESB 1}) can
3211 be used within structure definitions:
3228 This will ensure that the structure members are sensibly aligned
3229 relative to the base of the structure.
3231 A final caveat: \c{ALIGN} and \c{ALIGNB} work relative to the
3232 beginning of the \e{section}, not the beginning of the address space
3233 in the final executable. Aligning to a 16-byte boundary when the
3234 section you're in is only guaranteed to be aligned to a 4-byte
3235 boundary, for example, is a waste of effort. Again, NASM does not
3236 check that the section's alignment characteristics are sensible for
3237 the use of \c{ALIGN} or \c{ALIGNB}.
3240 \H{tasmcompat} \i{TASM Compatible Preprocessor Directives}
3242 The following preprocessor directives may only be used when TASM
3243 compatibility is turned on using the \c{-t} command line switch
3244 (This switch is described in \k{opt-t}.)
3246 \b\c{%arg} (see \k{arg})
3248 \b\c{%stacksize} (see \k{stacksize})
3250 \b\c{%local} (see \k{local})
3253 \S{arg} \i\c{%arg} Directive
3255 The \c{%arg} directive is used to simplify the handling of
3256 parameters passed on the stack. Stack based parameter passing
3257 is used by many high level languages, including C, C++ and Pascal.
3259 While NASM comes with macros which attempt to duplicate this
3260 functionality (see \k{16cmacro}), the syntax is not particularly
3261 convenient to use and is not TASM compatible. Here is an example
3262 which shows the use of \c{%arg} without any external macros:
3266 \c %push mycontext ; save the current context
3267 \c %stacksize large ; tell NASM to use bp
3268 \c %arg i:word, j_ptr:word
3275 \c %pop ; restore original context
3277 This is similar to the procedure defined in \k{16cmacro} and adds
3278 the value in i to the value pointed to by j_ptr and returns the
3279 sum in the ax register. See \k{pushpop} for an explanation of
3280 \c{push} and \c{pop} and the use of context stacks.
3283 \S{stacksize} \i\c{%stacksize} Directive
3285 The \c{%stacksize} directive is used in conjunction with the
3286 \c{%arg} (see \k{arg}) and the \c{%local} (see \k{local}) directives.
3287 It tells NASM the default size to use for subsequent \c{%arg} and
3288 \c{%local} directives. The \c{%stacksize} directive takes one
3289 required argument which is one of \c{flat}, \c{large} or \c{small}.
3293 This form causes NASM to use stack-based parameter addressing
3294 relative to \c{ebp} and it assumes that a near form of call was used
3295 to get to this label (i.e. that \c{eip} is on the stack).
3299 This form uses \c{bp} to do stack-based parameter addressing and
3300 assumes that a far form of call was used to get to this address
3301 (i.e. that \c{ip} and \c{cs} are on the stack).
3305 This form also uses \c{bp} to address stack parameters, but it is
3306 different from \c{large} because it also assumes that the old value
3307 of bp is pushed onto the stack (i.e. it expects an \c{ENTER}
3308 instruction). In other words, it expects that \c{bp}, \c{ip} and
3309 \c{cs} are on the top of the stack, underneath any local space which
3310 may have been allocated by \c{ENTER}. This form is probably most
3311 useful when used in combination with the \c{%local} directive
3315 \S{local} \i\c{%local} Directive
3317 The \c{%local} directive is used to simplify the use of local
3318 temporary stack variables allocated in a stack frame. Automatic
3319 local variables in C are an example of this kind of variable. The
3320 \c{%local} directive is most useful when used with the \c{%stacksize}
3321 (see \k{stacksize} and is also compatible with the \c{%arg} directive
3322 (see \k{arg}). It allows simplified reference to variables on the
3323 stack which have been allocated typically by using the \c{ENTER}
3325 \# (see \k{insENTER} for a description of that instruction).
3326 An example of its use is the following:
3330 \c %push mycontext ; save the current context
3331 \c %stacksize small ; tell NASM to use bp
3332 \c %assign %$localsize 0 ; see text for explanation
3333 \c %local old_ax:word, old_dx:word
3335 \c enter %$localsize,0 ; see text for explanation
3336 \c mov [old_ax],ax ; swap ax & bx
3337 \c mov [old_dx],dx ; and swap dx & cx
3342 \c leave ; restore old bp
3345 \c %pop ; restore original context
3347 The \c{%$localsize} variable is used internally by the
3348 \c{%local} directive and \e{must} be defined within the
3349 current context before the \c{%local} directive may be used.
3350 Failure to do so will result in one expression syntax error for
3351 each \c{%local} variable declared. It then may be used in
3352 the construction of an appropriately sized ENTER instruction
3353 as shown in the example.
3355 \H{otherpreproc} \i{Other Preprocessor Directives}
3357 NASM also has preprocessor directives which allow access to
3358 information from external sources. Currently they include:
3360 The following preprocessor directive is supported to allow NASM to
3361 correctly handle output of the cpp C language preprocessor.
3363 \b\c{%line} enables NAsM to correctly handle the output of the cpp
3364 C language preprocessor (see \k{line}).
3366 \b\c{%!} enables NASM to read in the value of an environment variable,
3367 which can then be used in your program (see \k{getenv}).
3369 \S{line} \i\c{%line} Directive
3371 The \c{%line} directive is used to notify NASM that the input line
3372 corresponds to a specific line number in another file. Typically
3373 this other file would be an original source file, with the current
3374 NASM input being the output of a pre-processor. The \c{%line}
3375 directive allows NASM to output messages which indicate the line
3376 number of the original source file, instead of the file that is being
3379 This preprocessor directive is not generally of use to programmers,
3380 by may be of interest to preprocessor authors. The usage of the
3381 \c{%line} preprocessor directive is as follows:
3383 \c %line nnn[+mmm] [filename]
3385 In this directive, \c{nnn} indentifies the line of the original source
3386 file which this line corresponds to. \c{mmm} is an optional parameter
3387 which specifies a line increment value; each line of the input file
3388 read in is considered to correspond to \c{mmm} lines of the original
3389 source file. Finally, \c{filename} is an optional parameter which
3390 specifies the file name of the original source file.
3392 After reading a \c{%line} preprocessor directive, NASM will report
3393 all file name and line numbers relative to the values specified
3397 \S{getenv} \i\c{%!}\c{<env>}: Read an environment variable.
3399 The \c{%!<env>} directive makes it possible to read the value of an
3400 environment variable at assembly time. This could, for example, be used
3401 to store the contents of an environment variable into a string, which
3402 could be used at some other point in your code.
3404 For example, suppose that you have an environment variable \c{FOO}, and
3405 you want the contents of \c{FOO} to be embedded in your program. You
3406 could do that as follows:
3408 \c %define FOO %!FOO
3411 \c tmpstr db quote FOO quote
3413 At the time of writing, this will generate an "unterminated string"
3414 warning at the time of defining "quote", and it will add a space
3415 before and after the string that is read in. I was unable to find
3416 a simple workaround (although a workaround can be created using a
3417 multi-line macro), so I believe that you will need to either learn how
3418 to create more complex macros, or allow for the extra spaces if you
3419 make use of this feature in that way.
3422 \C{directive} \i{Assembler Directives}
3424 NASM, though it attempts to avoid the bureaucracy of assemblers like
3425 MASM and TASM, is nevertheless forced to support a \e{few}
3426 directives. These are described in this chapter.
3428 NASM's directives come in two types: \I{user-level
3429 directives}\e{user-level} directives and \I{primitive
3430 directives}\e{primitive} directives. Typically, each directive has a
3431 user-level form and a primitive form. In almost all cases, we
3432 recommend that users use the user-level forms of the directives,
3433 which are implemented as macros which call the primitive forms.
3435 Primitive directives are enclosed in square brackets; user-level
3438 In addition to the universal directives described in this chapter,
3439 each object file format can optionally supply extra directives in
3440 order to control particular features of that file format. These
3441 \I{format-specific directives}\e{format-specific} directives are
3442 documented along with the formats that implement them, in \k{outfmt}.
3445 \H{bits} \i\c{BITS}: Specifying Target \i{Processor Mode}
3447 The \c{BITS} directive specifies whether NASM should generate code
3448 \I{16-bit mode, versus 32-bit mode}designed to run on a processor
3449 operating in 16-bit mode, 32-bit mode or 64-bit mode. The syntax is
3450 \c{BITS XX}, where XX is 16, 32 or 64.
3452 In most cases, you should not need to use \c{BITS} explicitly. The
3453 \c{aout}, \c{coff}, \c{elf}, \c{macho}, \c{win32} and \c{win64}
3454 object formats, which are designed for use in 32-bit or 64-bit
3455 operating systems, all cause NASM to select 32-bit or 64-bit mode,
3456 respectively, by default. The \c{obj} object format allows you
3457 to specify each segment you define as either \c{USE16} or \c{USE32},
3458 and NASM will set its operating mode accordingly, so the use of the
3459 \c{BITS} directive is once again unnecessary.
3461 The most likely reason for using the \c{BITS} directive is to write
3462 32-bit or 64-bit code in a flat binary file; this is because the \c{bin}
3463 output format defaults to 16-bit mode in anticipation of it being
3464 used most frequently to write DOS \c{.COM} programs, DOS \c{.SYS}
3465 device drivers and boot loader software.
3467 You do \e{not} need to specify \c{BITS 32} merely in order to use
3468 32-bit instructions in a 16-bit DOS program; if you do, the
3469 assembler will generate incorrect code because it will be writing
3470 code targeted at a 32-bit platform, to be run on a 16-bit one.
3472 When NASM is in \c{BITS 16} mode, instructions which use 32-bit
3473 data are prefixed with an 0x66 byte, and those referring to 32-bit
3474 addresses have an 0x67 prefix. In \c{BITS 32} mode, the reverse is
3475 true: 32-bit instructions require no prefixes, whereas instructions
3476 using 16-bit data need an 0x66 and those working on 16-bit addresses
3479 When NASM is in \c{BITS 64} mode, most instructions operate the same
3480 as they do for \c{BITS 32} mode. However, there are 8 more general and
3481 SSE registers, and 16-bit addressing is no longer supported.
3483 The default address size is 64 bits; 32-bit addressing can be selected
3484 with the 0x67 prefix. The default operand size is still 32 bits,
3485 however, and the 0x66 prefix selects 16-bit operand size. The \c{REX}
3486 prefix is used both to select 64-bit operand size, and to access the
3487 new registers. NASM automatically inserts REX prefixes when
3490 When the \c{REX} prefix is used, the processor does not know how to
3491 address the AH, BH, CH or DH (high 8-bit legacy) registers. Instead,
3492 it is possible to access the the low 8-bits of the SP, BP SI and DI
3493 registers as SPL, BPL, SIL and DIL, respectively; but only when the
3496 The \c{BITS} directive has an exactly equivalent primitive form,
3497 \c{[BITS 16]}, \c{[BITS 32]} and \c{[BITS 64]}. The user-level form is
3498 a macro which has no function other than to call the primitive form.
3500 Note that the space is neccessary, e.g. \c{BITS32} will \e{not} work!
3502 \S{USE16 & USE32} \i\c{USE16} & \i\c{USE32}: Aliases for BITS
3504 The `\c{USE16}' and `\c{USE32}' directives can be used in place of
3505 `\c{BITS 16}' and `\c{BITS 32}', for compatibility with other assemblers.
3508 \H{default} \i\c{DEFAULT}: Change the assembler defaults
3510 The \c{DEFAULT} directive changes the assembler defaults. Normally,
3511 NASM defaults to a mode where the programmer is expected to explicitly
3512 specify most features directly. However, this is occationally
3513 obnoxious, as the explicit form is pretty much the only one one wishes
3516 Currently, the only \c{DEFAULT} that is settable is whether or not
3517 registerless instructions in 64-bit mode are \c{RIP}-relative or not.
3518 By default, they are absolute unless overridden with the \i\c{REL}
3519 specifier (see \k{effaddr}). However, if \c{DEFAULT REL} is
3520 specified, \c{REL} is default, unless overridden with the \c{ABS}
3521 specifier, \e{except when used with an FS or GS segment override}.
3523 The special handling of \c{FS} and \c{GS} overrides are due to the
3524 fact that these registers are generally used as thread pointers or
3525 other special functions in 64-bit mode, and generating
3526 \c{RIP}-relative addresses would be extremely confusing.
3528 \c{DEFAULT REL} is disabled with \c{DEFAULT ABS}.
3530 \H{section} \i\c{SECTION} or \i\c{SEGMENT}: Changing and \i{Defining
3533 \I{changing sections}\I{switching between sections}The \c{SECTION}
3534 directive (\c{SEGMENT} is an exactly equivalent synonym) changes
3535 which section of the output file the code you write will be
3536 assembled into. In some object file formats, the number and names of
3537 sections are fixed; in others, the user may make up as many as they
3538 wish. Hence \c{SECTION} may sometimes give an error message, or may
3539 define a new section, if you try to switch to a section that does
3542 The Unix object formats, and the \c{bin} object format (but see
3543 \k{multisec}, all support
3544 the \i{standardized section names} \c{.text}, \c{.data} and \c{.bss}
3545 for the code, data and uninitialized-data sections. The \c{obj}
3546 format, by contrast, does not recognize these section names as being
3547 special, and indeed will strip off the leading period of any section
3551 \S{sectmac} The \i\c{__SECT__} Macro
3553 The \c{SECTION} directive is unusual in that its user-level form
3554 functions differently from its primitive form. The primitive form,
3555 \c{[SECTION xyz]}, simply switches the current target section to the
3556 one given. The user-level form, \c{SECTION xyz}, however, first
3557 defines the single-line macro \c{__SECT__} to be the primitive
3558 \c{[SECTION]} directive which it is about to issue, and then issues
3559 it. So the user-level directive
3563 expands to the two lines
3565 \c %define __SECT__ [SECTION .text]
3568 Users may find it useful to make use of this in their own macros.
3569 For example, the \c{writefile} macro defined in \k{mlmacgre} can be
3570 usefully rewritten in the following more sophisticated form:
3572 \c %macro writefile 2+
3582 \c mov cx,%%endstr-%%str
3589 This form of the macro, once passed a string to output, first
3590 switches temporarily to the data section of the file, using the
3591 primitive form of the \c{SECTION} directive so as not to modify
3592 \c{__SECT__}. It then declares its string in the data section, and
3593 then invokes \c{__SECT__} to switch back to \e{whichever} section
3594 the user was previously working in. It thus avoids the need, in the
3595 previous version of the macro, to include a \c{JMP} instruction to
3596 jump over the data, and also does not fail if, in a complicated
3597 \c{OBJ} format module, the user could potentially be assembling the
3598 code in any of several separate code sections.
3601 \H{absolute} \i\c{ABSOLUTE}: Defining Absolute Labels
3603 The \c{ABSOLUTE} directive can be thought of as an alternative form
3604 of \c{SECTION}: it causes the subsequent code to be directed at no
3605 physical section, but at the hypothetical section starting at the
3606 given absolute address. The only instructions you can use in this
3607 mode are the \c{RESB} family.
3609 \c{ABSOLUTE} is used as follows:
3617 This example describes a section of the PC BIOS data area, at
3618 segment address 0x40: the above code defines \c{kbuf_chr} to be
3619 0x1A, \c{kbuf_free} to be 0x1C, and \c{kbuf} to be 0x1E.
3621 The user-level form of \c{ABSOLUTE}, like that of \c{SECTION},
3622 redefines the \i\c{__SECT__} macro when it is invoked.
3624 \i\c{STRUC} and \i\c{ENDSTRUC} are defined as macros which use
3625 \c{ABSOLUTE} (and also \c{__SECT__}).
3627 \c{ABSOLUTE} doesn't have to take an absolute constant as an
3628 argument: it can take an expression (actually, a \i{critical
3629 expression}: see \k{crit}) and it can be a value in a segment. For
3630 example, a TSR can re-use its setup code as run-time BSS like this:
3632 \c org 100h ; it's a .COM program
3634 \c jmp setup ; setup code comes last
3636 \c ; the resident part of the TSR goes here
3638 \c ; now write the code that installs the TSR here
3642 \c runtimevar1 resw 1
3643 \c runtimevar2 resd 20
3647 This defines some variables `on top of' the setup code, so that
3648 after the setup has finished running, the space it took up can be
3649 re-used as data storage for the running TSR. The symbol `tsr_end'
3650 can be used to calculate the total size of the part of the TSR that
3651 needs to be made resident.
3654 \H{extern} \i\c{EXTERN}: \i{Importing Symbols} from Other Modules
3656 \c{EXTERN} is similar to the MASM directive \c{EXTRN} and the C
3657 keyword \c{extern}: it is used to declare a symbol which is not
3658 defined anywhere in the module being assembled, but is assumed to be
3659 defined in some other module and needs to be referred to by this
3660 one. Not every object-file format can support external variables:
3661 the \c{bin} format cannot.
3663 The \c{EXTERN} directive takes as many arguments as you like. Each
3664 argument is the name of a symbol:
3667 \c extern _sscanf,_fscanf
3669 Some object-file formats provide extra features to the \c{EXTERN}
3670 directive. In all cases, the extra features are used by suffixing a
3671 colon to the symbol name followed by object-format specific text.
3672 For example, the \c{obj} format allows you to declare that the
3673 default segment base of an external should be the group \c{dgroup}
3674 by means of the directive
3676 \c extern _variable:wrt dgroup
3678 The primitive form of \c{EXTERN} differs from the user-level form
3679 only in that it can take only one argument at a time: the support
3680 for multiple arguments is implemented at the preprocessor level.
3682 You can declare the same variable as \c{EXTERN} more than once: NASM
3683 will quietly ignore the second and later redeclarations. You can't
3684 declare a variable as \c{EXTERN} as well as something else, though.
3687 \H{global} \i\c{GLOBAL}: \i{Exporting Symbols} to Other Modules
3689 \c{GLOBAL} is the other end of \c{EXTERN}: if one module declares a
3690 symbol as \c{EXTERN} and refers to it, then in order to prevent
3691 linker errors, some other module must actually \e{define} the
3692 symbol and declare it as \c{GLOBAL}. Some assemblers use the name
3693 \i\c{PUBLIC} for this purpose.
3695 The \c{GLOBAL} directive applying to a symbol must appear \e{before}
3696 the definition of the symbol.
3698 \c{GLOBAL} uses the same syntax as \c{EXTERN}, except that it must
3699 refer to symbols which \e{are} defined in the same module as the
3700 \c{GLOBAL} directive. For example:
3706 \c{GLOBAL}, like \c{EXTERN}, allows object formats to define private
3707 extensions by means of a colon. The \c{elf} object format, for
3708 example, lets you specify whether global data items are functions or
3711 \c global hashlookup:function, hashtable:data
3713 Like \c{EXTERN}, the primitive form of \c{GLOBAL} differs from the
3714 user-level form only in that it can take only one argument at a
3718 \H{common} \i\c{COMMON}: Defining Common Data Areas
3720 The \c{COMMON} directive is used to declare \i\e{common variables}.
3721 A common variable is much like a global variable declared in the
3722 uninitialized data section, so that
3726 is similar in function to
3733 The difference is that if more than one module defines the same
3734 common variable, then at link time those variables will be
3735 \e{merged}, and references to \c{intvar} in all modules will point
3736 at the same piece of memory.
3738 Like \c{GLOBAL} and \c{EXTERN}, \c{COMMON} supports object-format
3739 specific extensions. For example, the \c{obj} format allows common
3740 variables to be NEAR or FAR, and the \c{elf} format allows you to
3741 specify the alignment requirements of a common variable:
3743 \c common commvar 4:near ; works in OBJ
3744 \c common intarray 100:4 ; works in ELF: 4 byte aligned
3746 Once again, like \c{EXTERN} and \c{GLOBAL}, the primitive form of
3747 \c{COMMON} differs from the user-level form only in that it can take
3748 only one argument at a time.
3751 \H{CPU} \i\c{CPU}: Defining CPU Dependencies
3753 The \i\c{CPU} directive restricts assembly to those instructions which
3754 are available on the specified CPU.
3758 \b\c{CPU 8086} Assemble only 8086 instruction set
3760 \b\c{CPU 186} Assemble instructions up to the 80186 instruction set
3762 \b\c{CPU 286} Assemble instructions up to the 286 instruction set
3764 \b\c{CPU 386} Assemble instructions up to the 386 instruction set
3766 \b\c{CPU 486} 486 instruction set
3768 \b\c{CPU 586} Pentium instruction set
3770 \b\c{CPU PENTIUM} Same as 586
3772 \b\c{CPU 686} P6 instruction set
3774 \b\c{CPU PPRO} Same as 686
3776 \b\c{CPU P2} Same as 686
3778 \b\c{CPU P3} Pentium III (Katmai) instruction sets
3780 \b\c{CPU KATMAI} Same as P3
3782 \b\c{CPU P4} Pentium 4 (Willamette) instruction set
3784 \b\c{CPU WILLAMETTE} Same as P4
3786 \b\c{CPU PRESCOTT} Prescott instruction set
3788 \b\c{CPU X64} x86-64 (x64/AMD64/EM64T) instruction set
3790 \b\c{CPU IA64} IA64 CPU (in x86 mode) instruction set
3792 All options are case insensitive. All instructions will be selected
3793 only if they apply to the selected CPU or lower. By default, all
3794 instructions are available.
3797 \C{outfmt} \i{Output Formats}
3799 NASM is a portable assembler, designed to be able to compile on any
3800 ANSI C-supporting platform and produce output to run on a variety of
3801 Intel x86 operating systems. For this reason, it has a large number
3802 of available output formats, selected using the \i\c{-f} option on
3803 the NASM \i{command line}. Each of these formats, along with its
3804 extensions to the base NASM syntax, is detailed in this chapter.
3806 As stated in \k{opt-o}, NASM chooses a \i{default name} for your
3807 output file based on the input file name and the chosen output
3808 format. This will be generated by removing the \i{extension}
3809 (\c{.asm}, \c{.s}, or whatever you like to use) from the input file
3810 name, and substituting an extension defined by the output format.
3811 The extensions are given with each format below.
3814 \H{binfmt} \i\c{bin}: \i{Flat-Form Binary}\I{pure binary} Output
3816 The \c{bin} format does not produce object files: it generates
3817 nothing in the output file except the code you wrote. Such `pure
3818 binary' files are used by \i{MS-DOS}: \i\c{.COM} executables and
3819 \i\c{.SYS} device drivers are pure binary files. Pure binary output
3820 is also useful for \i{operating system} and \i{boot loader}
3823 The \c{bin} format supports \i{multiple section names}. For details of
3824 how nasm handles sections in the \c{bin} format, see \k{multisec}.
3826 Using the \c{bin} format puts NASM by default into 16-bit mode (see
3827 \k{bits}). In order to use \c{bin} to write 32-bit or 64-bit code,
3828 such as an OS kernel, you need to explicitly issue the \I\c{BITS}\c{BITS 32}
3829 or \I\c{BITS}\c{BITS 64} directive.
3831 \c{bin} has no default output file name extension: instead, it
3832 leaves your file name as it is once the original extension has been
3833 removed. Thus, the default is for NASM to assemble \c{binprog.asm}
3834 into a binary file called \c{binprog}.
3837 \S{org} \i\c{ORG}: Binary File \i{Program Origin}
3839 The \c{bin} format provides an additional directive to the list
3840 given in \k{directive}: \c{ORG}. The function of the \c{ORG}
3841 directive is to specify the origin address which NASM will assume
3842 the program begins at when it is loaded into memory.
3844 For example, the following code will generate the longword
3851 Unlike the \c{ORG} directive provided by MASM-compatible assemblers,
3852 which allows you to jump around in the object file and overwrite
3853 code you have already generated, NASM's \c{ORG} does exactly what
3854 the directive says: \e{origin}. Its sole function is to specify one
3855 offset which is added to all internal address references within the
3856 section; it does not permit any of the trickery that MASM's version
3857 does. See \k{proborg} for further comments.
3860 \S{binseg} \c{bin} Extensions to the \c{SECTION}
3861 Directive\I{SECTION, bin extensions to}
3863 The \c{bin} output format extends the \c{SECTION} (or \c{SEGMENT})
3864 directive to allow you to specify the alignment requirements of
3865 segments. This is done by appending the \i\c{ALIGN} qualifier to the
3866 end of the section-definition line. For example,
3868 \c section .data align=16
3870 switches to the section \c{.data} and also specifies that it must be
3871 aligned on a 16-byte boundary.
3873 The parameter to \c{ALIGN} specifies how many low bits of the
3874 section start address must be forced to zero. The alignment value
3875 given may be any power of two.\I{section alignment, in
3876 bin}\I{segment alignment, in bin}\I{alignment, in bin sections}
3879 \S{multisec} \i\c{Multisection}\I{bin, multisection} support for the BIN format.
3881 The \c{bin} format allows the use of multiple sections, of arbitrary names,
3882 besides the "known" \c{.text}, \c{.data}, and \c{.bss} names.
3884 \b Sections may be designated \i\c{progbits} or \i\c{nobits}. Default
3885 is \c{progbits} (except \c{.bss}, which defaults to \c{nobits},
3888 \b Sections can be aligned at a specified boundary following the previous
3889 section with \c{align=}, or at an arbitrary byte-granular position with
3892 \b Sections can be given a virtual start address, which will be used
3893 for the calculation of all memory references within that section
3896 \b Sections can be ordered using \i\c{follows=}\c{<section>} or
3897 \i\c{vfollows=}\c{<section>} as an alternative to specifying an explicit
3900 \b Arguments to \c{org}, \c{start}, \c{vstart}, and \c{align=} are
3901 critical expressions. See \k{crit}. E.g. \c{align=(1 << ALIGN_SHIFT)}
3902 - \c{ALIGN_SHIFT} must be defined before it is used here.
3904 \b Any code which comes before an explicit \c{SECTION} directive
3905 is directed by default into the \c{.text} section.
3907 \b If an \c{ORG} statement is not given, \c{ORG 0} is used
3910 \b The \c{.bss} section will be placed after the last \c{progbits}
3911 section, unless \c{start=}, \c{vstart=}, \c{follows=}, or \c{vfollows=}
3914 \b All sections are aligned on dword boundaries, unless a different
3915 alignment has been specified.
3917 \b Sections may not overlap.
3919 \b Nasm creates the \c{section.<secname>.start} for each section,
3920 which may be used in your code.
3922 \S{map}\i{Map files}
3924 Map files can be generated in \c{-f bin} format by means of the \c{[map]}
3925 option. Map types of \c{all} (default), \c{brief}, \c{sections}, \c{segments},
3926 or \c{symbols} may be specified. Output may be directed to \c{stdout}
3927 (default), \c{stderr}, or a specified file. E.g.
3928 \c{[map symbols myfile.map]}. No "user form" exists, the square
3929 brackets must be used.
3932 \H{objfmt} \i\c{obj}: \i{Microsoft OMF}\I{OMF} Object Files
3934 The \c{obj} file format (NASM calls it \c{obj} rather than \c{omf}
3935 for historical reasons) is the one produced by \i{MASM} and
3936 \i{TASM}, which is typically fed to 16-bit DOS linkers to produce
3937 \i\c{.EXE} files. It is also the format used by \i{OS/2}.
3939 \c{obj} provides a default output file-name extension of \c{.obj}.
3941 \c{obj} is not exclusively a 16-bit format, though: NASM has full
3942 support for the 32-bit extensions to the format. In particular,
3943 32-bit \c{obj} format files are used by \i{Borland's Win32
3944 compilers}, instead of using Microsoft's newer \i\c{win32} object
3947 The \c{obj} format does not define any special segment names: you
3948 can call your segments anything you like. Typical names for segments
3949 in \c{obj} format files are \c{CODE}, \c{DATA} and \c{BSS}.
3951 If your source file contains code before specifying an explicit
3952 \c{SEGMENT} directive, then NASM will invent its own segment called
3953 \i\c{__NASMDEFSEG} for you.
3955 When you define a segment in an \c{obj} file, NASM defines the
3956 segment name as a symbol as well, so that you can access the segment
3957 address of the segment. So, for example:
3966 \c mov ax,data ; get segment address of data
3967 \c mov ds,ax ; and move it into DS
3968 \c inc word [dvar] ; now this reference will work
3971 The \c{obj} format also enables the use of the \i\c{SEG} and
3972 \i\c{WRT} operators, so that you can write code which does things
3977 \c mov ax,seg foo ; get preferred segment of foo
3979 \c mov ax,data ; a different segment
3981 \c mov ax,[ds:foo] ; this accesses `foo'
3982 \c mov [es:foo wrt data],bx ; so does this
3985 \S{objseg} \c{obj} Extensions to the \c{SEGMENT}
3986 Directive\I{SEGMENT, obj extensions to}
3988 The \c{obj} output format extends the \c{SEGMENT} (or \c{SECTION})
3989 directive to allow you to specify various properties of the segment
3990 you are defining. This is done by appending extra qualifiers to the
3991 end of the segment-definition line. For example,
3993 \c segment code private align=16
3995 defines the segment \c{code}, but also declares it to be a private
3996 segment, and requires that the portion of it described in this code
3997 module must be aligned on a 16-byte boundary.
3999 The available qualifiers are:
4001 \b \i\c{PRIVATE}, \i\c{PUBLIC}, \i\c{COMMON} and \i\c{STACK} specify
4002 the combination characteristics of the segment. \c{PRIVATE} segments
4003 do not get combined with any others by the linker; \c{PUBLIC} and
4004 \c{STACK} segments get concatenated together at link time; and
4005 \c{COMMON} segments all get overlaid on top of each other rather
4006 than stuck end-to-end.
4008 \b \i\c{ALIGN} is used, as shown above, to specify how many low bits
4009 of the segment start address must be forced to zero. The alignment
4010 value given may be any power of two from 1 to 4096; in reality, the
4011 only values supported are 1, 2, 4, 16, 256 and 4096, so if 8 is
4012 specified it will be rounded up to 16, and 32, 64 and 128 will all
4013 be rounded up to 256, and so on. Note that alignment to 4096-byte
4014 boundaries is a \i{PharLap} extension to the format and may not be
4015 supported by all linkers.\I{section alignment, in OBJ}\I{segment
4016 alignment, in OBJ}\I{alignment, in OBJ sections}
4018 \b \i\c{CLASS} can be used to specify the segment class; this feature
4019 indicates to the linker that segments of the same class should be
4020 placed near each other in the output file. The class name can be any
4021 word, e.g. \c{CLASS=CODE}.
4023 \b \i\c{OVERLAY}, like \c{CLASS}, is specified with an arbitrary word
4024 as an argument, and provides overlay information to an
4025 overlay-capable linker.
4027 \b Segments can be declared as \i\c{USE16} or \i\c{USE32}, which has
4028 the effect of recording the choice in the object file and also
4029 ensuring that NASM's default assembly mode when assembling in that
4030 segment is 16-bit or 32-bit respectively.
4032 \b When writing \i{OS/2} object files, you should declare 32-bit
4033 segments as \i\c{FLAT}, which causes the default segment base for
4034 anything in the segment to be the special group \c{FLAT}, and also
4035 defines the group if it is not already defined.
4037 \b The \c{obj} file format also allows segments to be declared as
4038 having a pre-defined absolute segment address, although no linkers
4039 are currently known to make sensible use of this feature;
4040 nevertheless, NASM allows you to declare a segment such as
4041 \c{SEGMENT SCREEN ABSOLUTE=0xB800} if you need to. The \i\c{ABSOLUTE}
4042 and \c{ALIGN} keywords are mutually exclusive.
4044 NASM's default segment attributes are \c{PUBLIC}, \c{ALIGN=1}, no
4045 class, no overlay, and \c{USE16}.
4048 \S{group} \i\c{GROUP}: Defining Groups of Segments\I{segments, groups of}
4050 The \c{obj} format also allows segments to be grouped, so that a
4051 single segment register can be used to refer to all the segments in
4052 a group. NASM therefore supplies the \c{GROUP} directive, whereby
4061 \c ; some uninitialized data
4063 \c group dgroup data bss
4065 which will define a group called \c{dgroup} to contain the segments
4066 \c{data} and \c{bss}. Like \c{SEGMENT}, \c{GROUP} causes the group
4067 name to be defined as a symbol, so that you can refer to a variable
4068 \c{var} in the \c{data} segment as \c{var wrt data} or as \c{var wrt
4069 dgroup}, depending on which segment value is currently in your
4072 If you just refer to \c{var}, however, and \c{var} is declared in a
4073 segment which is part of a group, then NASM will default to giving
4074 you the offset of \c{var} from the beginning of the \e{group}, not
4075 the \e{segment}. Therefore \c{SEG var}, also, will return the group
4076 base rather than the segment base.
4078 NASM will allow a segment to be part of more than one group, but
4079 will generate a warning if you do this. Variables declared in a
4080 segment which is part of more than one group will default to being
4081 relative to the first group that was defined to contain the segment.
4083 A group does not have to contain any segments; you can still make
4084 \c{WRT} references to a group which does not contain the variable
4085 you are referring to. OS/2, for example, defines the special group
4086 \c{FLAT} with no segments in it.
4089 \S{uppercase} \i\c{UPPERCASE}: Disabling Case Sensitivity in Output
4091 Although NASM itself is \i{case sensitive}, some OMF linkers are
4092 not; therefore it can be useful for NASM to output single-case
4093 object files. The \c{UPPERCASE} format-specific directive causes all
4094 segment, group and symbol names that are written to the object file
4095 to be forced to upper case just before being written. Within a
4096 source file, NASM is still case-sensitive; but the object file can
4097 be written entirely in upper case if desired.
4099 \c{UPPERCASE} is used alone on a line; it requires no parameters.
4102 \S{import} \i\c{IMPORT}: Importing DLL Symbols\I{DLL symbols,
4103 importing}\I{symbols, importing from DLLs}
4105 The \c{IMPORT} format-specific directive defines a symbol to be
4106 imported from a DLL, for use if you are writing a DLL's \i{import
4107 library} in NASM. You still need to declare the symbol as \c{EXTERN}
4108 as well as using the \c{IMPORT} directive.
4110 The \c{IMPORT} directive takes two required parameters, separated by
4111 white space, which are (respectively) the name of the symbol you
4112 wish to import and the name of the library you wish to import it
4115 \c import WSAStartup wsock32.dll
4117 A third optional parameter gives the name by which the symbol is
4118 known in the library you are importing it from, in case this is not
4119 the same as the name you wish the symbol to be known by to your code
4120 once you have imported it. For example:
4122 \c import asyncsel wsock32.dll WSAAsyncSelect
4125 \S{export} \i\c{EXPORT}: Exporting DLL Symbols\I{DLL symbols,
4126 exporting}\I{symbols, exporting from DLLs}
4128 The \c{EXPORT} format-specific directive defines a global symbol to
4129 be exported as a DLL symbol, for use if you are writing a DLL in
4130 NASM. You still need to declare the symbol as \c{GLOBAL} as well as
4131 using the \c{EXPORT} directive.
4133 \c{EXPORT} takes one required parameter, which is the name of the
4134 symbol you wish to export, as it was defined in your source file. An
4135 optional second parameter (separated by white space from the first)
4136 gives the \e{external} name of the symbol: the name by which you
4137 wish the symbol to be known to programs using the DLL. If this name
4138 is the same as the internal name, you may leave the second parameter
4141 Further parameters can be given to define attributes of the exported
4142 symbol. These parameters, like the second, are separated by white
4143 space. If further parameters are given, the external name must also
4144 be specified, even if it is the same as the internal name. The
4145 available attributes are:
4147 \b \c{resident} indicates that the exported name is to be kept
4148 resident by the system loader. This is an optimisation for
4149 frequently used symbols imported by name.
4151 \b \c{nodata} indicates that the exported symbol is a function which
4152 does not make use of any initialized data.
4154 \b \c{parm=NNN}, where \c{NNN} is an integer, sets the number of
4155 parameter words for the case in which the symbol is a call gate
4156 between 32-bit and 16-bit segments.
4158 \b An attribute which is just a number indicates that the symbol
4159 should be exported with an identifying number (ordinal), and gives
4165 \c export myfunc TheRealMoreFormalLookingFunctionName
4166 \c export myfunc myfunc 1234 ; export by ordinal
4167 \c export myfunc myfunc resident parm=23 nodata
4170 \S{dotdotstart} \i\c{..start}: Defining the \i{Program Entry
4173 \c{OMF} linkers require exactly one of the object files being linked to
4174 define the program entry point, where execution will begin when the
4175 program is run. If the object file that defines the entry point is
4176 assembled using NASM, you specify the entry point by declaring the
4177 special symbol \c{..start} at the point where you wish execution to
4181 \S{objextern} \c{obj} Extensions to the \c{EXTERN}
4182 Directive\I{EXTERN, obj extensions to}
4184 If you declare an external symbol with the directive
4188 then references such as \c{mov ax,foo} will give you the offset of
4189 \c{foo} from its preferred segment base (as specified in whichever
4190 module \c{foo} is actually defined in). So to access the contents of
4191 \c{foo} you will usually need to do something like
4193 \c mov ax,seg foo ; get preferred segment base
4194 \c mov es,ax ; move it into ES
4195 \c mov ax,[es:foo] ; and use offset `foo' from it
4197 This is a little unwieldy, particularly if you know that an external
4198 is going to be accessible from a given segment or group, say
4199 \c{dgroup}. So if \c{DS} already contained \c{dgroup}, you could
4202 \c mov ax,[foo wrt dgroup]
4204 However, having to type this every time you want to access \c{foo}
4205 can be a pain; so NASM allows you to declare \c{foo} in the
4208 \c extern foo:wrt dgroup
4210 This form causes NASM to pretend that the preferred segment base of
4211 \c{foo} is in fact \c{dgroup}; so the expression \c{seg foo} will
4212 now return \c{dgroup}, and the expression \c{foo} is equivalent to
4215 This \I{default-WRT mechanism}default-\c{WRT} mechanism can be used
4216 to make externals appear to be relative to any group or segment in
4217 your program. It can also be applied to common variables: see
4221 \S{objcommon} \c{obj} Extensions to the \c{COMMON}
4222 Directive\I{COMMON, obj extensions to}
4224 The \c{obj} format allows common variables to be either near\I{near
4225 common variables} or far\I{far common variables}; NASM allows you to
4226 specify which your variables should be by the use of the syntax
4228 \c common nearvar 2:near ; `nearvar' is a near common
4229 \c common farvar 10:far ; and `farvar' is far
4231 Far common variables may be greater in size than 64Kb, and so the
4232 OMF specification says that they are declared as a number of
4233 \e{elements} of a given size. So a 10-byte far common variable could
4234 be declared as ten one-byte elements, five two-byte elements, two
4235 five-byte elements or one ten-byte element.
4237 Some \c{OMF} linkers require the \I{element size, in common
4238 variables}\I{common variables, element size}element size, as well as
4239 the variable size, to match when resolving common variables declared
4240 in more than one module. Therefore NASM must allow you to specify
4241 the element size on your far common variables. This is done by the
4244 \c common c_5by2 10:far 5 ; two five-byte elements
4245 \c common c_2by5 10:far 2 ; five two-byte elements
4247 If no element size is specified, the default is 1. Also, the \c{FAR}
4248 keyword is not required when an element size is specified, since
4249 only far commons may have element sizes at all. So the above
4250 declarations could equivalently be
4252 \c common c_5by2 10:5 ; two five-byte elements
4253 \c common c_2by5 10:2 ; five two-byte elements
4255 In addition to these extensions, the \c{COMMON} directive in \c{obj}
4256 also supports default-\c{WRT} specification like \c{EXTERN} does
4257 (explained in \k{objextern}). So you can also declare things like
4259 \c common foo 10:wrt dgroup
4260 \c common bar 16:far 2:wrt data
4261 \c common baz 24:wrt data:6
4264 \H{win32fmt} \i\c{win32}: Microsoft Win32 Object Files
4266 The \c{win32} output format generates Microsoft Win32 object files,
4267 suitable for passing to Microsoft linkers such as \i{Visual C++}.
4268 Note that Borland Win32 compilers do not use this format, but use
4269 \c{obj} instead (see \k{objfmt}).
4271 \c{win32} provides a default output file-name extension of \c{.obj}.
4273 Note that although Microsoft say that Win32 object files follow the
4274 \c{COFF} (Common Object File Format) standard, the object files produced
4275 by Microsoft Win32 compilers are not compatible with COFF linkers
4276 such as DJGPP's, and vice versa. This is due to a difference of
4277 opinion over the precise semantics of PC-relative relocations. To
4278 produce COFF files suitable for DJGPP, use NASM's \c{coff} output
4279 format; conversely, the \c{coff} format does not produce object
4280 files that Win32 linkers can generate correct output from.
4283 \S{win32sect} \c{win32} Extensions to the \c{SECTION}
4284 Directive\I{SECTION, win32 extensions to}
4286 Like the \c{obj} format, \c{win32} allows you to specify additional
4287 information on the \c{SECTION} directive line, to control the type
4288 and properties of sections you declare. Section types and properties
4289 are generated automatically by NASM for the \i{standard section names}
4290 \c{.text}, \c{.data} and \c{.bss}, but may still be overridden by
4293 The available qualifiers are:
4295 \b \c{code}, or equivalently \c{text}, defines the section to be a
4296 code section. This marks the section as readable and executable, but
4297 not writable, and also indicates to the linker that the type of the
4300 \b \c{data} and \c{bss} define the section to be a data section,
4301 analogously to \c{code}. Data sections are marked as readable and
4302 writable, but not executable. \c{data} declares an initialized data
4303 section, whereas \c{bss} declares an uninitialized data section.
4305 \b \c{rdata} declares an initialized data section that is readable
4306 but not writable. Microsoft compilers use this section to place
4309 \b \c{info} defines the section to be an \i{informational section},
4310 which is not included in the executable file by the linker, but may
4311 (for example) pass information \e{to} the linker. For example,
4312 declaring an \c{info}-type section called \i\c{.drectve} causes the
4313 linker to interpret the contents of the section as command-line
4316 \b \c{align=}, used with a trailing number as in \c{obj}, gives the
4317 \I{section alignment, in win32}\I{alignment, in win32
4318 sections}alignment requirements of the section. The maximum you may
4319 specify is 64: the Win32 object file format contains no means to
4320 request a greater section alignment than this. If alignment is not
4321 explicitly specified, the defaults are 16-byte alignment for code
4322 sections, 8-byte alignment for rdata sections and 4-byte alignment
4323 for data (and BSS) sections.
4324 Informational sections get a default alignment of 1 byte (no
4325 alignment), though the value does not matter.
4327 The defaults assumed by NASM if you do not specify the above
4330 \c section .text code align=16
4331 \c section .data data align=4
4332 \c section .rdata rdata align=8
4333 \c section .bss bss align=4
4335 Any other section name is treated by default like \c{.text}.
4338 \H{win64fmt} \i\c{win64}: Microsoft Win64 Object Files
4340 The \c{win64} output format generates Microsoft Win64 object files,
4341 which is nearly 100% indentical to the \c{win32} object format (\k{win32fmt})
4342 with the exception that it is meant to target 64-bit code and the x86-64
4343 platform altogether. This object file is used exactly the same as the \c{win32}
4344 object format (\k{win32fmt}), in NASM, with regard to this exception.
4347 \H{cofffmt} \i\c{coff}: \i{Common Object File Format}
4349 The \c{coff} output type produces \c{COFF} object files suitable for
4350 linking with the \i{DJGPP} linker.
4352 \c{coff} provides a default output file-name extension of \c{.o}.
4354 The \c{coff} format supports the same extensions to the \c{SECTION}
4355 directive as \c{win32} does, except that the \c{align} qualifier and
4356 the \c{info} section type are not supported.
4358 \H{machofmt} \i\c{macho}: \i{Mach Object File Format}
4360 The \c{macho} output type produces \c{Mach-O} object files suitable for
4361 linking with the \i{Mac OSX} linker.
4363 \c{macho} provides a default output file-name extension of \c{.o}.
4365 \H{elffmt} \i\c{elf, elf32, and elf64}: \I{ELF}\I{linux, elf}\i{Executable and Linkable
4366 Format} Object Files
4368 The \c{elf32} and \c{elf64} output formats generate \c{ELF32 and ELF64} (Executable and Linkable Format) object files, as used by Linux as well as \i{Unix System V},
4369 including \i{Solaris x86}, \i{UnixWare} and \i{SCO Unix}. \c{elf}
4370 provides a default output file-name extension of \c{.o}. \c{elf} is a synonym for \c{elf32}.
4373 \S{elfsect} \c{elf} Extensions to the \c{SECTION}
4374 Directive\I{SECTION, elf extensions to}
4376 Like the \c{obj} format, \c{elf} allows you to specify additional
4377 information on the \c{SECTION} directive line, to control the type
4378 and properties of sections you declare. Section types and properties
4379 are generated automatically by NASM for the \i{standard section
4380 names} \i\c{.text}, \i\c{.data} and \i\c{.bss}, but may still be
4381 overridden by these qualifiers.
4383 The available qualifiers are:
4385 \b \i\c{alloc} defines the section to be one which is loaded into
4386 memory when the program is run. \i\c{noalloc} defines it to be one
4387 which is not, such as an informational or comment section.
4389 \b \i\c{exec} defines the section to be one which should have execute
4390 permission when the program is run. \i\c{noexec} defines it as one
4393 \b \i\c{write} defines the section to be one which should be writable
4394 when the program is run. \i\c{nowrite} defines it as one which should
4397 \b \i\c{progbits} defines the section to be one with explicit contents
4398 stored in the object file: an ordinary code or data section, for
4399 example, \i\c{nobits} defines the section to be one with no explicit
4400 contents given, such as a BSS section.
4402 \b \c{align=}, used with a trailing number as in \c{obj}, gives the
4403 \I{section alignment, in elf}\I{alignment, in elf sections}alignment
4404 requirements of the section.
4406 The defaults assumed by NASM if you do not specify the above
4409 \c section .text progbits alloc exec nowrite align=16
4410 \c section .rodata progbits alloc noexec nowrite align=4
4411 \c section .data progbits alloc noexec write align=4
4412 \c section .bss nobits alloc noexec write align=4
4413 \c section other progbits alloc noexec nowrite align=1
4415 (Any section name other than \c{.text}, \c{.rodata}, \c{.data} and
4416 \c{.bss} is treated by default like \c{other} in the above code.)
4419 \S{elfwrt} \i{Position-Independent Code}\I{PIC}: \c{elf} Special
4420 Symbols and \i\c{WRT}
4422 The \c{ELF} specification contains enough features to allow
4423 position-independent code (PIC) to be written, which makes \i{ELF
4424 shared libraries} very flexible. However, it also means NASM has to
4425 be able to generate a variety of strange relocation types in ELF
4426 object files, if it is to be an assembler which can write PIC.
4428 Since \c{ELF} does not support segment-base references, the \c{WRT}
4429 operator is not used for its normal purpose; therefore NASM's
4430 \c{elf} output format makes use of \c{WRT} for a different purpose,
4431 namely the PIC-specific \I{relocations, PIC-specific}relocation
4434 \c{elf} defines five special symbols which you can use as the
4435 right-hand side of the \c{WRT} operator to obtain PIC relocation
4436 types. They are \i\c{..gotpc}, \i\c{..gotoff}, \i\c{..got},
4437 \i\c{..plt} and \i\c{..sym}. Their functions are summarized here:
4439 \b Referring to the symbol marking the global offset table base
4440 using \c{wrt ..gotpc} will end up giving the distance from the
4441 beginning of the current section to the global offset table.
4442 (\i\c{_GLOBAL_OFFSET_TABLE_} is the standard symbol name used to
4443 refer to the \i{GOT}.) So you would then need to add \i\c{$$} to the
4444 result to get the real address of the GOT.
4446 \b Referring to a location in one of your own sections using \c{wrt
4447 ..gotoff} will give the distance from the beginning of the GOT to
4448 the specified location, so that adding on the address of the GOT
4449 would give the real address of the location you wanted.
4451 \b Referring to an external or global symbol using \c{wrt ..got}
4452 causes the linker to build an entry \e{in} the GOT containing the
4453 address of the symbol, and the reference gives the distance from the
4454 beginning of the GOT to the entry; so you can add on the address of
4455 the GOT, load from the resulting address, and end up with the
4456 address of the symbol.
4458 \b Referring to a procedure name using \c{wrt ..plt} causes the
4459 linker to build a \i{procedure linkage table} entry for the symbol,
4460 and the reference gives the address of the \i{PLT} entry. You can
4461 only use this in contexts which would generate a PC-relative
4462 relocation normally (i.e. as the destination for \c{CALL} or
4463 \c{JMP}), since ELF contains no relocation type to refer to PLT
4466 \b Referring to a symbol name using \c{wrt ..sym} causes NASM to
4467 write an ordinary relocation, but instead of making the relocation
4468 relative to the start of the section and then adding on the offset
4469 to the symbol, it will write a relocation record aimed directly at
4470 the symbol in question. The distinction is a necessary one due to a
4471 peculiarity of the dynamic linker.
4473 A fuller explanation of how to use these relocation types to write
4474 shared libraries entirely in NASM is given in \k{picdll}.
4477 \S{elfglob} \c{elf} Extensions to the \c{GLOBAL} Directive\I{GLOBAL,
4478 elf extensions to}\I{GLOBAL, aoutb extensions to}
4480 \c{ELF} object files can contain more information about a global symbol
4481 than just its address: they can contain the \I{symbol sizes,
4482 specifying}\I{size, of symbols}size of the symbol and its \I{symbol
4483 types, specifying}\I{type, of symbols}type as well. These are not
4484 merely debugger conveniences, but are actually necessary when the
4485 program being written is a \i{shared library}. NASM therefore
4486 supports some extensions to the \c{GLOBAL} directive, allowing you
4487 to specify these features.
4489 You can specify whether a global variable is a function or a data
4490 object by suffixing the name with a colon and the word
4491 \i\c{function} or \i\c{data}. (\i\c{object} is a synonym for
4492 \c{data}.) For example:
4494 \c global hashlookup:function, hashtable:data
4496 exports the global symbol \c{hashlookup} as a function and
4497 \c{hashtable} as a data object.
4499 Optionally, you can control the ELF visibility of the symbol. Just
4500 add one of the visibility keywords: \i\c{default}, \i\c{internal},
4501 \i\c{hidden}, or \i\c{protected}. The default is \i\c{default} of
4502 course. For example, to make \c{hashlookup} hidden:
4504 \c global hashlookup:function hidden
4506 You can also specify the size of the data associated with the
4507 symbol, as a numeric expression (which may involve labels, and even
4508 forward references) after the type specifier. Like this:
4510 \c global hashtable:data (hashtable.end - hashtable)
4513 \c db this,that,theother ; some data here
4516 This makes NASM automatically calculate the length of the table and
4517 place that information into the \c{ELF} symbol table.
4519 Declaring the type and size of global symbols is necessary when
4520 writing shared library code. For more information, see
4524 \S{elfcomm} \c{elf} Extensions to the \c{COMMON} Directive
4525 \I{COMMON, elf extensions to}
4527 \c{ELF} also allows you to specify alignment requirements \I{common
4528 variables, alignment in elf}\I{alignment, of elf common variables}on
4529 common variables. This is done by putting a number (which must be a
4530 power of two) after the name and size of the common variable,
4531 separated (as usual) by a colon. For example, an array of
4532 doublewords would benefit from 4-byte alignment:
4534 \c common dwordarray 128:4
4536 This declares the total size of the array to be 128 bytes, and
4537 requires that it be aligned on a 4-byte boundary.
4540 \S{elf16} 16-bit code and ELF
4541 \I{ELF, 16-bit code and}
4543 The \c{ELF32} specification doesn't provide relocations for 8- and
4544 16-bit values, but the GNU \c{ld} linker adds these as an extension.
4545 NASM can generate GNU-compatible relocations, to allow 16-bit code to
4546 be linked as ELF using GNU \c{ld}. If NASM is used with the
4547 \c{-w+gnu-elf-extensions} option, a warning is issued when one of
4548 these relocations is generated.
4550 \H{aoutfmt} \i\c{aout}: Linux \I{a.out, Linux version}\I{linux, a.out}\c{a.out} Object Files
4552 The \c{aout} format generates \c{a.out} object files, in the form used
4553 by early Linux systems (current Linux systems use ELF, see
4554 \k{elffmt}.) These differ from other \c{a.out} object files in that
4555 the magic number in the first four bytes of the file is
4556 different; also, some implementations of \c{a.out}, for example
4557 NetBSD's, support position-independent code, which Linux's
4558 implementation does not.
4560 \c{a.out} provides a default output file-name extension of \c{.o}.
4562 \c{a.out} is a very simple object format. It supports no special
4563 directives, no special symbols, no use of \c{SEG} or \c{WRT}, and no
4564 extensions to any standard directives. It supports only the three
4565 \i{standard section names} \i\c{.text}, \i\c{.data} and \i\c{.bss}.
4568 \H{aoutfmt} \i\c{aoutb}: \i{NetBSD}/\i{FreeBSD}/\i{OpenBSD}
4569 \I{a.out, BSD version}\c{a.out} Object Files
4571 The \c{aoutb} format generates \c{a.out} object files, in the form
4572 used by the various free \c{BSD Unix} clones, \c{NetBSD}, \c{FreeBSD}
4573 and \c{OpenBSD}. For simple object files, this object format is exactly
4574 the same as \c{aout} except for the magic number in the first four bytes
4575 of the file. However, the \c{aoutb} format supports
4576 \I{PIC}\i{position-independent code} in the same way as the \c{elf}
4577 format, so you can use it to write \c{BSD} \i{shared libraries}.
4579 \c{aoutb} provides a default output file-name extension of \c{.o}.
4581 \c{aoutb} supports no special directives, no special symbols, and
4582 only the three \i{standard section names} \i\c{.text}, \i\c{.data}
4583 and \i\c{.bss}. However, it also supports the same use of \i\c{WRT} as
4584 \c{elf} does, to provide position-independent code relocation types.
4585 See \k{elfwrt} for full documentation of this feature.
4587 \c{aoutb} also supports the same extensions to the \c{GLOBAL}
4588 directive as \c{elf} does: see \k{elfglob} for documentation of
4592 \H{as86fmt} \c{as86}: \i{Minix}/Linux\I{linux, as86} \i\c{as86} Object Files
4594 The Minix/Linux 16-bit assembler \c{as86} has its own non-standard
4595 object file format. Although its companion linker \i\c{ld86} produces
4596 something close to ordinary \c{a.out} binaries as output, the object
4597 file format used to communicate between \c{as86} and \c{ld86} is not
4600 NASM supports this format, just in case it is useful, as \c{as86}.
4601 \c{as86} provides a default output file-name extension of \c{.o}.
4603 \c{as86} is a very simple object format (from the NASM user's point
4604 of view). It supports no special directives, no special symbols, no
4605 use of \c{SEG} or \c{WRT}, and no extensions to any standard
4606 directives. It supports only the three \i{standard section names}
4607 \i\c{.text}, \i\c{.data} and \i\c{.bss}.
4610 \H{rdffmt} \I{RDOFF}\i\c{rdf}: \i{Relocatable Dynamic Object File
4613 The \c{rdf} output format produces \c{RDOFF} object files. \c{RDOFF}
4614 (Relocatable Dynamic Object File Format) is a home-grown object-file
4615 format, designed alongside NASM itself and reflecting in its file
4616 format the internal structure of the assembler.
4618 \c{RDOFF} is not used by any well-known operating systems. Those
4619 writing their own systems, however, may well wish to use \c{RDOFF}
4620 as their object format, on the grounds that it is designed primarily
4621 for simplicity and contains very little file-header bureaucracy.
4623 The Unix NASM archive, and the DOS archive which includes sources,
4624 both contain an \I{rdoff subdirectory}\c{rdoff} subdirectory holding
4625 a set of RDOFF utilities: an RDF linker, an \c{RDF} static-library
4626 manager, an RDF file dump utility, and a program which will load and
4627 execute an RDF executable under Linux.
4629 \c{rdf} supports only the \i{standard section names} \i\c{.text},
4630 \i\c{.data} and \i\c{.bss}.
4633 \S{rdflib} Requiring a Library: The \i\c{LIBRARY} Directive
4635 \c{RDOFF} contains a mechanism for an object file to demand a given
4636 library to be linked to the module, either at load time or run time.
4637 This is done by the \c{LIBRARY} directive, which takes one argument
4638 which is the name of the module:
4640 \c library mylib.rdl
4643 \S{rdfmod} Specifying a Module Name: The \i\c{MODULE} Directive
4645 Special \c{RDOFF} header record is used to store the name of the module.
4646 It can be used, for example, by run-time loader to perform dynamic
4647 linking. \c{MODULE} directive takes one argument which is the name
4652 Note that when you statically link modules and tell linker to strip
4653 the symbols from output file, all module names will be stripped too.
4654 To avoid it, you should start module names with \I{$, prefix}\c{$}, like:
4656 \c module $kernel.core
4659 \S{rdfglob} \c{rdf} Extensions to the \c{GLOBAL} directive\I{GLOBAL,
4662 \c{RDOFF} global symbols can contain additional information needed by
4663 the static linker. You can mark a global symbol as exported, thus
4664 telling the linker do not strip it from target executable or library
4665 file. Like in \c{ELF}, you can also specify whether an exported symbol
4666 is a procedure (function) or data object.
4668 Suffixing the name with a colon and the word \i\c{export} you make the
4671 \c global sys_open:export
4673 To specify that exported symbol is a procedure (function), you add the
4674 word \i\c{proc} or \i\c{function} after declaration:
4676 \c global sys_open:export proc
4678 Similarly, to specify exported data object, add the word \i\c{data}
4679 or \i\c{object} to the directive:
4681 \c global kernel_ticks:export data
4684 \S{rdfimpt} \c{rdf} Extensions to the \c{EXTERN} directive\I{EXTERN,
4687 By default the \c{EXTERN} directive in \c{RDOFF} declares a "pure external"
4688 symbol (i.e. the static linker will complain if such a symbol is not resolved).
4689 To declare an "imported" symbol, which must be resolved later during a dynamic
4690 linking phase, \c{RDOFF} offers an additional \c{import} modifier. As in
4691 \c{GLOBAL}, you can also specify whether an imported symbol is a procedure
4692 (function) or data object. For example:
4695 \c extern _open:import
4696 \c extern _printf:import proc
4697 \c extern _errno:import data
4699 Here the directive \c{LIBRARY} is also included, which gives the dynamic linker
4700 a hint as to where to find requested symbols.
4703 \H{dbgfmt} \i\c{dbg}: Debugging Format
4705 The \c{dbg} output format is not built into NASM in the default
4706 configuration. If you are building your own NASM executable from the
4707 sources, you can define \i\c{OF_DBG} in \c{outform.h} or on the
4708 compiler command line, and obtain the \c{dbg} output format.
4710 The \c{dbg} format does not output an object file as such; instead,
4711 it outputs a text file which contains a complete list of all the
4712 transactions between the main body of NASM and the output-format
4713 back end module. It is primarily intended to aid people who want to
4714 write their own output drivers, so that they can get a clearer idea
4715 of the various requests the main program makes of the output driver,
4716 and in what order they happen.
4718 For simple files, one can easily use the \c{dbg} format like this:
4720 \c nasm -f dbg filename.asm
4722 which will generate a diagnostic file called \c{filename.dbg}.
4723 However, this will not work well on files which were designed for a
4724 different object format, because each object format defines its own
4725 macros (usually user-level forms of directives), and those macros
4726 will not be defined in the \c{dbg} format. Therefore it can be
4727 useful to run NASM twice, in order to do the preprocessing with the
4728 native object format selected:
4730 \c nasm -e -f rdf -o rdfprog.i rdfprog.asm
4731 \c nasm -a -f dbg rdfprog.i
4733 This preprocesses \c{rdfprog.asm} into \c{rdfprog.i}, keeping the
4734 \c{rdf} object format selected in order to make sure RDF special
4735 directives are converted into primitive form correctly. Then the
4736 preprocessed source is fed through the \c{dbg} format to generate
4737 the final diagnostic output.
4739 This workaround will still typically not work for programs intended
4740 for \c{obj} format, because the \c{obj} \c{SEGMENT} and \c{GROUP}
4741 directives have side effects of defining the segment and group names
4742 as symbols; \c{dbg} will not do this, so the program will not
4743 assemble. You will have to work around that by defining the symbols
4744 yourself (using \c{EXTERN}, for example) if you really need to get a
4745 \c{dbg} trace of an \c{obj}-specific source file.
4747 \c{dbg} accepts any section name and any directives at all, and logs
4748 them all to its output file.
4751 \C{16bit} Writing 16-bit Code (DOS, Windows 3/3.1)
4753 This chapter attempts to cover some of the common issues encountered
4754 when writing 16-bit code to run under \c{MS-DOS} or \c{Windows 3.x}. It
4755 covers how to link programs to produce \c{.EXE} or \c{.COM} files,
4756 how to write \c{.SYS} device drivers, and how to interface assembly
4757 language code with 16-bit C compilers and with Borland Pascal.
4760 \H{exefiles} Producing \i\c{.EXE} Files
4762 Any large program written under DOS needs to be built as a \c{.EXE}
4763 file: only \c{.EXE} files have the necessary internal structure
4764 required to span more than one 64K segment. \i{Windows} programs,
4765 also, have to be built as \c{.EXE} files, since Windows does not
4766 support the \c{.COM} format.
4768 In general, you generate \c{.EXE} files by using the \c{obj} output
4769 format to produce one or more \i\c{.OBJ} files, and then linking
4770 them together using a linker. However, NASM also supports the direct
4771 generation of simple DOS \c{.EXE} files using the \c{bin} output
4772 format (by using \c{DB} and \c{DW} to construct the \c{.EXE} file
4773 header), and a macro package is supplied to do this. Thanks to
4774 Yann Guidon for contributing the code for this.
4776 NASM may also support \c{.EXE} natively as another output format in
4780 \S{objexe} Using the \c{obj} Format To Generate \c{.EXE} Files
4782 This section describes the usual method of generating \c{.EXE} files
4783 by linking \c{.OBJ} files together.
4785 Most 16-bit programming language packages come with a suitable
4786 linker; if you have none of these, there is a free linker called
4787 \i{VAL}\I{linker, free}, available in \c{LZH} archive format from
4788 \W{ftp://x2ftp.oulu.fi/pub/msdos/programming/lang/}\i\c{x2ftp.oulu.fi}.
4789 An LZH archiver can be found at
4790 \W{ftp://ftp.simtel.net/pub/simtelnet/msdos/arcers}\i\c{ftp.simtel.net}.
4791 There is another `free' linker (though this one doesn't come with
4792 sources) called \i{FREELINK}, available from
4793 \W{http://www.pcorner.com/tpc/old/3-101.html}\i\c{www.pcorner.com}.
4794 A third, \i\c{djlink}, written by DJ Delorie, is available at
4795 \W{http://www.delorie.com/djgpp/16bit/djlink/}\i\c{www.delorie.com}.
4796 A fourth linker, \i\c{ALINK}, written by Anthony A.J. Williams, is
4797 available at \W{http://alink.sourceforge.net}\i\c{alink.sourceforge.net}.
4799 When linking several \c{.OBJ} files into a \c{.EXE} file, you should
4800 ensure that exactly one of them has a start point defined (using the
4801 \I{program entry point}\i\c{..start} special symbol defined by the
4802 \c{obj} format: see \k{dotdotstart}). If no module defines a start
4803 point, the linker will not know what value to give the entry-point
4804 field in the output file header; if more than one defines a start
4805 point, the linker will not know \e{which} value to use.
4807 An example of a NASM source file which can be assembled to a
4808 \c{.OBJ} file and linked on its own to a \c{.EXE} is given here. It
4809 demonstrates the basic principles of defining a stack, initialising
4810 the segment registers, and declaring a start point. This file is
4811 also provided in the \I{test subdirectory}\c{test} subdirectory of
4812 the NASM archives, under the name \c{objexe.asm}.
4823 This initial piece of code sets up \c{DS} to point to the data
4824 segment, and initializes \c{SS} and \c{SP} to point to the top of
4825 the provided stack. Notice that interrupts are implicitly disabled
4826 for one instruction after a move into \c{SS}, precisely for this
4827 situation, so that there's no chance of an interrupt occurring
4828 between the loads of \c{SS} and \c{SP} and not having a stack to
4831 Note also that the special symbol \c{..start} is defined at the
4832 beginning of this code, which means that will be the entry point
4833 into the resulting executable file.
4839 The above is the main program: load \c{DS:DX} with a pointer to the
4840 greeting message (\c{hello} is implicitly relative to the segment
4841 \c{data}, which was loaded into \c{DS} in the setup code, so the
4842 full pointer is valid), and call the DOS print-string function.
4847 This terminates the program using another DOS system call.
4851 \c hello: db 'hello, world', 13, 10, '$'
4853 The data segment contains the string we want to display.
4855 \c segment stack stack
4859 The above code declares a stack segment containing 64 bytes of
4860 uninitialized stack space, and points \c{stacktop} at the top of it.
4861 The directive \c{segment stack stack} defines a segment \e{called}
4862 \c{stack}, and also of \e{type} \c{STACK}. The latter is not
4863 necessary to the correct running of the program, but linkers are
4864 likely to issue warnings or errors if your program has no segment of
4867 The above file, when assembled into a \c{.OBJ} file, will link on
4868 its own to a valid \c{.EXE} file, which when run will print `hello,
4869 world' and then exit.
4872 \S{binexe} Using the \c{bin} Format To Generate \c{.EXE} Files
4874 The \c{.EXE} file format is simple enough that it's possible to
4875 build a \c{.EXE} file by writing a pure-binary program and sticking
4876 a 32-byte header on the front. This header is simple enough that it
4877 can be generated using \c{DB} and \c{DW} commands by NASM itself, so
4878 that you can use the \c{bin} output format to directly generate
4881 Included in the NASM archives, in the \I{misc subdirectory}\c{misc}
4882 subdirectory, is a file \i\c{exebin.mac} of macros. It defines three
4883 macros: \i\c{EXE_begin}, \i\c{EXE_stack} and \i\c{EXE_end}.
4885 To produce a \c{.EXE} file using this method, you should start by
4886 using \c{%include} to load the \c{exebin.mac} macro package into
4887 your source file. You should then issue the \c{EXE_begin} macro call
4888 (which takes no arguments) to generate the file header data. Then
4889 write code as normal for the \c{bin} format - you can use all three
4890 standard sections \c{.text}, \c{.data} and \c{.bss}. At the end of
4891 the file you should call the \c{EXE_end} macro (again, no arguments),
4892 which defines some symbols to mark section sizes, and these symbols
4893 are referred to in the header code generated by \c{EXE_begin}.
4895 In this model, the code you end up writing starts at \c{0x100}, just
4896 like a \c{.COM} file - in fact, if you strip off the 32-byte header
4897 from the resulting \c{.EXE} file, you will have a valid \c{.COM}
4898 program. All the segment bases are the same, so you are limited to a
4899 64K program, again just like a \c{.COM} file. Note that an \c{ORG}
4900 directive is issued by the \c{EXE_begin} macro, so you should not
4901 explicitly issue one of your own.
4903 You can't directly refer to your segment base value, unfortunately,
4904 since this would require a relocation in the header, and things
4905 would get a lot more complicated. So you should get your segment
4906 base by copying it out of \c{CS} instead.
4908 On entry to your \c{.EXE} file, \c{SS:SP} are already set up to
4909 point to the top of a 2Kb stack. You can adjust the default stack
4910 size of 2Kb by calling the \c{EXE_stack} macro. For example, to
4911 change the stack size of your program to 64 bytes, you would call
4914 A sample program which generates a \c{.EXE} file in this way is
4915 given in the \c{test} subdirectory of the NASM archive, as
4919 \H{comfiles} Producing \i\c{.COM} Files
4921 While large DOS programs must be written as \c{.EXE} files, small
4922 ones are often better written as \c{.COM} files. \c{.COM} files are
4923 pure binary, and therefore most easily produced using the \c{bin}
4927 \S{combinfmt} Using the \c{bin} Format To Generate \c{.COM} Files
4929 \c{.COM} files expect to be loaded at offset \c{100h} into their
4930 segment (though the segment may change). Execution then begins at
4931 \I\c{ORG}\c{100h}, i.e. right at the start of the program. So to
4932 write a \c{.COM} program, you would create a source file looking
4940 \c ; put your code here
4944 \c ; put data items here
4948 \c ; put uninitialized data here
4950 The \c{bin} format puts the \c{.text} section first in the file, so
4951 you can declare data or BSS items before beginning to write code if
4952 you want to and the code will still end up at the front of the file
4955 The BSS (uninitialized data) section does not take up space in the
4956 \c{.COM} file itself: instead, addresses of BSS items are resolved
4957 to point at space beyond the end of the file, on the grounds that
4958 this will be free memory when the program is run. Therefore you
4959 should not rely on your BSS being initialized to all zeros when you
4962 To assemble the above program, you should use a command line like
4964 \c nasm myprog.asm -fbin -o myprog.com
4966 The \c{bin} format would produce a file called \c{myprog} if no
4967 explicit output file name were specified, so you have to override it
4968 and give the desired file name.
4971 \S{comobjfmt} Using the \c{obj} Format To Generate \c{.COM} Files
4973 If you are writing a \c{.COM} program as more than one module, you
4974 may wish to assemble several \c{.OBJ} files and link them together
4975 into a \c{.COM} program. You can do this, provided you have a linker
4976 capable of outputting \c{.COM} files directly (\i{TLINK} does this),
4977 or alternatively a converter program such as \i\c{EXE2BIN} to
4978 transform the \c{.EXE} file output from the linker into a \c{.COM}
4981 If you do this, you need to take care of several things:
4983 \b The first object file containing code should start its code
4984 segment with a line like \c{RESB 100h}. This is to ensure that the
4985 code begins at offset \c{100h} relative to the beginning of the code
4986 segment, so that the linker or converter program does not have to
4987 adjust address references within the file when generating the
4988 \c{.COM} file. Other assemblers use an \i\c{ORG} directive for this
4989 purpose, but \c{ORG} in NASM is a format-specific directive to the
4990 \c{bin} output format, and does not mean the same thing as it does
4991 in MASM-compatible assemblers.
4993 \b You don't need to define a stack segment.
4995 \b All your segments should be in the same group, so that every time
4996 your code or data references a symbol offset, all offsets are
4997 relative to the same segment base. This is because, when a \c{.COM}
4998 file is loaded, all the segment registers contain the same value.
5001 \H{sysfiles} Producing \i\c{.SYS} Files
5003 \i{MS-DOS device drivers} - \c{.SYS} files - are pure binary files,
5004 similar to \c{.COM} files, except that they start at origin zero
5005 rather than \c{100h}. Therefore, if you are writing a device driver
5006 using the \c{bin} format, you do not need the \c{ORG} directive,
5007 since the default origin for \c{bin} is zero. Similarly, if you are
5008 using \c{obj}, you do not need the \c{RESB 100h} at the start of
5011 \c{.SYS} files start with a header structure, containing pointers to
5012 the various routines inside the driver which do the work. This
5013 structure should be defined at the start of the code segment, even
5014 though it is not actually code.
5016 For more information on the format of \c{.SYS} files, and the data
5017 which has to go in the header structure, a list of books is given in
5018 the Frequently Asked Questions list for the newsgroup
5019 \W{news:comp.os.msdos.programmer}\i\c{comp.os.msdos.programmer}.
5022 \H{16c} Interfacing to 16-bit C Programs
5024 This section covers the basics of writing assembly routines that
5025 call, or are called from, C programs. To do this, you would
5026 typically write an assembly module as a \c{.OBJ} file, and link it
5027 with your C modules to produce a \i{mixed-language program}.
5030 \S{16cunder} External Symbol Names
5032 \I{C symbol names}\I{underscore, in C symbols}C compilers have the
5033 convention that the names of all global symbols (functions or data)
5034 they define are formed by prefixing an underscore to the name as it
5035 appears in the C program. So, for example, the function a C
5036 programmer thinks of as \c{printf} appears to an assembly language
5037 programmer as \c{_printf}. This means that in your assembly
5038 programs, you can define symbols without a leading underscore, and
5039 not have to worry about name clashes with C symbols.
5041 If you find the underscores inconvenient, you can define macros to
5042 replace the \c{GLOBAL} and \c{EXTERN} directives as follows:
5058 (These forms of the macros only take one argument at a time; a
5059 \c{%rep} construct could solve this.)
5061 If you then declare an external like this:
5065 then the macro will expand it as
5068 \c %define printf _printf
5070 Thereafter, you can reference \c{printf} as if it was a symbol, and
5071 the preprocessor will put the leading underscore on where necessary.
5073 The \c{cglobal} macro works similarly. You must use \c{cglobal}
5074 before defining the symbol in question, but you would have had to do
5075 that anyway if you used \c{GLOBAL}.
5077 Also see \k{opt-pfix}.
5079 \S{16cmodels} \i{Memory Models}
5081 NASM contains no mechanism to support the various C memory models
5082 directly; you have to keep track yourself of which one you are
5083 writing for. This means you have to keep track of the following
5086 \b In models using a single code segment (tiny, small and compact),
5087 functions are near. This means that function pointers, when stored
5088 in data segments or pushed on the stack as function arguments, are
5089 16 bits long and contain only an offset field (the \c{CS} register
5090 never changes its value, and always gives the segment part of the
5091 full function address), and that functions are called using ordinary
5092 near \c{CALL} instructions and return using \c{RETN} (which, in
5093 NASM, is synonymous with \c{RET} anyway). This means both that you
5094 should write your own routines to return with \c{RETN}, and that you
5095 should call external C routines with near \c{CALL} instructions.
5097 \b In models using more than one code segment (medium, large and
5098 huge), functions are far. This means that function pointers are 32
5099 bits long (consisting of a 16-bit offset followed by a 16-bit
5100 segment), and that functions are called using \c{CALL FAR} (or
5101 \c{CALL seg:offset}) and return using \c{RETF}. Again, you should
5102 therefore write your own routines to return with \c{RETF} and use
5103 \c{CALL FAR} to call external routines.
5105 \b In models using a single data segment (tiny, small and medium),
5106 data pointers are 16 bits long, containing only an offset field (the
5107 \c{DS} register doesn't change its value, and always gives the
5108 segment part of the full data item address).
5110 \b In models using more than one data segment (compact, large and
5111 huge), data pointers are 32 bits long, consisting of a 16-bit offset
5112 followed by a 16-bit segment. You should still be careful not to
5113 modify \c{DS} in your routines without restoring it afterwards, but
5114 \c{ES} is free for you to use to access the contents of 32-bit data
5115 pointers you are passed.
5117 \b The huge memory model allows single data items to exceed 64K in
5118 size. In all other memory models, you can access the whole of a data
5119 item just by doing arithmetic on the offset field of the pointer you
5120 are given, whether a segment field is present or not; in huge model,
5121 you have to be more careful of your pointer arithmetic.
5123 \b In most memory models, there is a \e{default} data segment, whose
5124 segment address is kept in \c{DS} throughout the program. This data
5125 segment is typically the same segment as the stack, kept in \c{SS},
5126 so that functions' local variables (which are stored on the stack)
5127 and global data items can both be accessed easily without changing
5128 \c{DS}. Particularly large data items are typically stored in other
5129 segments. However, some memory models (though not the standard
5130 ones, usually) allow the assumption that \c{SS} and \c{DS} hold the
5131 same value to be removed. Be careful about functions' local
5132 variables in this latter case.
5134 In models with a single code segment, the segment is called
5135 \i\c{_TEXT}, so your code segment must also go by this name in order
5136 to be linked into the same place as the main code segment. In models
5137 with a single data segment, or with a default data segment, it is
5141 \S{16cfunc} Function Definitions and Function Calls
5143 \I{functions, C calling convention}The \i{C calling convention} in
5144 16-bit programs is as follows. In the following description, the
5145 words \e{caller} and \e{callee} are used to denote the function
5146 doing the calling and the function which gets called.
5148 \b The caller pushes the function's parameters on the stack, one
5149 after another, in reverse order (right to left, so that the first
5150 argument specified to the function is pushed last).
5152 \b The caller then executes a \c{CALL} instruction to pass control
5153 to the callee. This \c{CALL} is either near or far depending on the
5156 \b The callee receives control, and typically (although this is not
5157 actually necessary, in functions which do not need to access their
5158 parameters) starts by saving the value of \c{SP} in \c{BP} so as to
5159 be able to use \c{BP} as a base pointer to find its parameters on
5160 the stack. However, the caller was probably doing this too, so part
5161 of the calling convention states that \c{BP} must be preserved by
5162 any C function. Hence the callee, if it is going to set up \c{BP} as
5163 a \i\e{frame pointer}, must push the previous value first.
5165 \b The callee may then access its parameters relative to \c{BP}.
5166 The word at \c{[BP]} holds the previous value of \c{BP} as it was
5167 pushed; the next word, at \c{[BP+2]}, holds the offset part of the
5168 return address, pushed implicitly by \c{CALL}. In a small-model
5169 (near) function, the parameters start after that, at \c{[BP+4]}; in
5170 a large-model (far) function, the segment part of the return address
5171 lives at \c{[BP+4]}, and the parameters begin at \c{[BP+6]}. The
5172 leftmost parameter of the function, since it was pushed last, is
5173 accessible at this offset from \c{BP}; the others follow, at
5174 successively greater offsets. Thus, in a function such as \c{printf}
5175 which takes a variable number of parameters, the pushing of the
5176 parameters in reverse order means that the function knows where to
5177 find its first parameter, which tells it the number and type of the
5180 \b The callee may also wish to decrease \c{SP} further, so as to
5181 allocate space on the stack for local variables, which will then be
5182 accessible at negative offsets from \c{BP}.
5184 \b The callee, if it wishes to return a value to the caller, should
5185 leave the value in \c{AL}, \c{AX} or \c{DX:AX} depending on the size
5186 of the value. Floating-point results are sometimes (depending on the
5187 compiler) returned in \c{ST0}.
5189 \b Once the callee has finished processing, it restores \c{SP} from
5190 \c{BP} if it had allocated local stack space, then pops the previous
5191 value of \c{BP}, and returns via \c{RETN} or \c{RETF} depending on
5194 \b When the caller regains control from the callee, the function
5195 parameters are still on the stack, so it typically adds an immediate
5196 constant to \c{SP} to remove them (instead of executing a number of
5197 slow \c{POP} instructions). Thus, if a function is accidentally
5198 called with the wrong number of parameters due to a prototype
5199 mismatch, the stack will still be returned to a sensible state since
5200 the caller, which \e{knows} how many parameters it pushed, does the
5203 It is instructive to compare this calling convention with that for
5204 Pascal programs (described in \k{16bpfunc}). Pascal has a simpler
5205 convention, since no functions have variable numbers of parameters.
5206 Therefore the callee knows how many parameters it should have been
5207 passed, and is able to deallocate them from the stack itself by
5208 passing an immediate argument to the \c{RET} or \c{RETF}
5209 instruction, so the caller does not have to do it. Also, the
5210 parameters are pushed in left-to-right order, not right-to-left,
5211 which means that a compiler can give better guarantees about
5212 sequence points without performance suffering.
5214 Thus, you would define a function in C style in the following way.
5215 The following example is for small model:
5222 \c sub sp,0x40 ; 64 bytes of local stack space
5223 \c mov bx,[bp+4] ; first parameter to function
5227 \c mov sp,bp ; undo "sub sp,0x40" above
5231 For a large-model function, you would replace \c{RET} by \c{RETF},
5232 and look for the first parameter at \c{[BP+6]} instead of
5233 \c{[BP+4]}. Of course, if one of the parameters is a pointer, then
5234 the offsets of \e{subsequent} parameters will change depending on
5235 the memory model as well: far pointers take up four bytes on the
5236 stack when passed as a parameter, whereas near pointers take up two.
5238 At the other end of the process, to call a C function from your
5239 assembly code, you would do something like this:
5243 \c ; and then, further down...
5245 \c push word [myint] ; one of my integer variables
5246 \c push word mystring ; pointer into my data segment
5248 \c add sp,byte 4 ; `byte' saves space
5250 \c ; then those data items...
5255 \c mystring db 'This number -> %d <- should be 1234',10,0
5257 This piece of code is the small-model assembly equivalent of the C
5260 \c int myint = 1234;
5261 \c printf("This number -> %d <- should be 1234\n", myint);
5263 In large model, the function-call code might look more like this. In
5264 this example, it is assumed that \c{DS} already holds the segment
5265 base of the segment \c{_DATA}. If not, you would have to initialize
5268 \c push word [myint]
5269 \c push word seg mystring ; Now push the segment, and...
5270 \c push word mystring ; ... offset of "mystring"
5274 The integer value still takes up one word on the stack, since large
5275 model does not affect the size of the \c{int} data type. The first
5276 argument (pushed last) to \c{printf}, however, is a data pointer,
5277 and therefore has to contain a segment and offset part. The segment
5278 should be stored second in memory, and therefore must be pushed
5279 first. (Of course, \c{PUSH DS} would have been a shorter instruction
5280 than \c{PUSH WORD SEG mystring}, if \c{DS} was set up as the above
5281 example assumed.) Then the actual call becomes a far call, since
5282 functions expect far calls in large model; and \c{SP} has to be
5283 increased by 6 rather than 4 afterwards to make up for the extra
5287 \S{16cdata} Accessing Data Items
5289 To get at the contents of C variables, or to declare variables which
5290 C can access, you need only declare the names as \c{GLOBAL} or
5291 \c{EXTERN}. (Again, the names require leading underscores, as stated
5292 in \k{16cunder}.) Thus, a C variable declared as \c{int i} can be
5293 accessed from assembler as
5299 And to declare your own integer variable which C programs can access
5300 as \c{extern int j}, you do this (making sure you are assembling in
5301 the \c{_DATA} segment, if necessary):
5307 To access a C array, you need to know the size of the components of
5308 the array. For example, \c{int} variables are two bytes long, so if
5309 a C program declares an array as \c{int a[10]}, you can access
5310 \c{a[3]} by coding \c{mov ax,[_a+6]}. (The byte offset 6 is obtained
5311 by multiplying the desired array index, 3, by the size of the array
5312 element, 2.) The sizes of the C base types in 16-bit compilers are:
5313 1 for \c{char}, 2 for \c{short} and \c{int}, 4 for \c{long} and
5314 \c{float}, and 8 for \c{double}.
5316 To access a C \i{data structure}, you need to know the offset from
5317 the base of the structure to the field you are interested in. You
5318 can either do this by converting the C structure definition into a
5319 NASM structure definition (using \i\c{STRUC}), or by calculating the
5320 one offset and using just that.
5322 To do either of these, you should read your C compiler's manual to
5323 find out how it organizes data structures. NASM gives no special
5324 alignment to structure members in its own \c{STRUC} macro, so you
5325 have to specify alignment yourself if the C compiler generates it.
5326 Typically, you might find that a structure like
5333 might be four bytes long rather than three, since the \c{int} field
5334 would be aligned to a two-byte boundary. However, this sort of
5335 feature tends to be a configurable option in the C compiler, either
5336 using command-line options or \c{#pragma} lines, so you have to find
5337 out how your own compiler does it.
5340 \S{16cmacro} \i\c{c16.mac}: Helper Macros for the 16-bit C Interface
5342 Included in the NASM archives, in the \I{misc subdirectory}\c{misc}
5343 directory, is a file \c{c16.mac} of macros. It defines three macros:
5344 \i\c{proc}, \i\c{arg} and \i\c{endproc}. These are intended to be
5345 used for C-style procedure definitions, and they automate a lot of
5346 the work involved in keeping track of the calling convention.
5348 (An alternative, TASM compatible form of \c{arg} is also now built
5349 into NASM's preprocessor. See \k{tasmcompat} for details.)
5351 An example of an assembly function using the macro set is given
5358 \c mov ax,[bp + %$i]
5359 \c mov bx,[bp + %$j]
5364 This defines \c{_nearproc} to be a procedure taking two arguments,
5365 the first (\c{i}) an integer and the second (\c{j}) a pointer to an
5366 integer. It returns \c{i + *j}.
5368 Note that the \c{arg} macro has an \c{EQU} as the first line of its
5369 expansion, and since the label before the macro call gets prepended
5370 to the first line of the expanded macro, the \c{EQU} works, defining
5371 \c{%$i} to be an offset from \c{BP}. A context-local variable is
5372 used, local to the context pushed by the \c{proc} macro and popped
5373 by the \c{endproc} macro, so that the same argument name can be used
5374 in later procedures. Of course, you don't \e{have} to do that.
5376 The macro set produces code for near functions (tiny, small and
5377 compact-model code) by default. You can have it generate far
5378 functions (medium, large and huge-model code) by means of coding
5379 \I\c{FARCODE}\c{%define FARCODE}. This changes the kind of return
5380 instruction generated by \c{endproc}, and also changes the starting
5381 point for the argument offsets. The macro set contains no intrinsic
5382 dependency on whether data pointers are far or not.
5384 \c{arg} can take an optional parameter, giving the size of the
5385 argument. If no size is given, 2 is assumed, since it is likely that
5386 many function parameters will be of type \c{int}.
5388 The large-model equivalent of the above function would look like this:
5396 \c mov ax,[bp + %$i]
5397 \c mov bx,[bp + %$j]
5398 \c mov es,[bp + %$j + 2]
5403 This makes use of the argument to the \c{arg} macro to define a
5404 parameter of size 4, because \c{j} is now a far pointer. When we
5405 load from \c{j}, we must load a segment and an offset.
5408 \H{16bp} Interfacing to \i{Borland Pascal} Programs
5410 Interfacing to Borland Pascal programs is similar in concept to
5411 interfacing to 16-bit C programs. The differences are:
5413 \b The leading underscore required for interfacing to C programs is
5414 not required for Pascal.
5416 \b The memory model is always large: functions are far, data
5417 pointers are far, and no data item can be more than 64K long.
5418 (Actually, some functions are near, but only those functions that
5419 are local to a Pascal unit and never called from outside it. All
5420 assembly functions that Pascal calls, and all Pascal functions that
5421 assembly routines are able to call, are far.) However, all static
5422 data declared in a Pascal program goes into the default data
5423 segment, which is the one whose segment address will be in \c{DS}
5424 when control is passed to your assembly code. The only things that
5425 do not live in the default data segment are local variables (they
5426 live in the stack segment) and dynamically allocated variables. All
5427 data \e{pointers}, however, are far.
5429 \b The function calling convention is different - described below.
5431 \b Some data types, such as strings, are stored differently.
5433 \b There are restrictions on the segment names you are allowed to
5434 use - Borland Pascal will ignore code or data declared in a segment
5435 it doesn't like the name of. The restrictions are described below.
5438 \S{16bpfunc} The Pascal Calling Convention
5440 \I{functions, Pascal calling convention}\I{Pascal calling
5441 convention}The 16-bit Pascal calling convention is as follows. In
5442 the following description, the words \e{caller} and \e{callee} are
5443 used to denote the function doing the calling and the function which
5446 \b The caller pushes the function's parameters on the stack, one
5447 after another, in normal order (left to right, so that the first
5448 argument specified to the function is pushed first).
5450 \b The caller then executes a far \c{CALL} instruction to pass
5451 control to the callee.
5453 \b The callee receives control, and typically (although this is not
5454 actually necessary, in functions which do not need to access their
5455 parameters) starts by saving the value of \c{SP} in \c{BP} so as to
5456 be able to use \c{BP} as a base pointer to find its parameters on
5457 the stack. However, the caller was probably doing this too, so part
5458 of the calling convention states that \c{BP} must be preserved by
5459 any function. Hence the callee, if it is going to set up \c{BP} as a
5460 \i{frame pointer}, must push the previous value first.
5462 \b The callee may then access its parameters relative to \c{BP}.
5463 The word at \c{[BP]} holds the previous value of \c{BP} as it was
5464 pushed. The next word, at \c{[BP+2]}, holds the offset part of the
5465 return address, and the next one at \c{[BP+4]} the segment part. The
5466 parameters begin at \c{[BP+6]}. The rightmost parameter of the
5467 function, since it was pushed last, is accessible at this offset
5468 from \c{BP}; the others follow, at successively greater offsets.
5470 \b The callee may also wish to decrease \c{SP} further, so as to
5471 allocate space on the stack for local variables, which will then be
5472 accessible at negative offsets from \c{BP}.
5474 \b The callee, if it wishes to return a value to the caller, should
5475 leave the value in \c{AL}, \c{AX} or \c{DX:AX} depending on the size
5476 of the value. Floating-point results are returned in \c{ST0}.
5477 Results of type \c{Real} (Borland's own custom floating-point data
5478 type, not handled directly by the FPU) are returned in \c{DX:BX:AX}.
5479 To return a result of type \c{String}, the caller pushes a pointer
5480 to a temporary string before pushing the parameters, and the callee
5481 places the returned string value at that location. The pointer is
5482 not a parameter, and should not be removed from the stack by the
5483 \c{RETF} instruction.
5485 \b Once the callee has finished processing, it restores \c{SP} from
5486 \c{BP} if it had allocated local stack space, then pops the previous
5487 value of \c{BP}, and returns via \c{RETF}. It uses the form of
5488 \c{RETF} with an immediate parameter, giving the number of bytes
5489 taken up by the parameters on the stack. This causes the parameters
5490 to be removed from the stack as a side effect of the return
5493 \b When the caller regains control from the callee, the function
5494 parameters have already been removed from the stack, so it needs to
5497 Thus, you would define a function in Pascal style, taking two
5498 \c{Integer}-type parameters, in the following way:
5504 \c sub sp,0x40 ; 64 bytes of local stack space
5505 \c mov bx,[bp+8] ; first parameter to function
5506 \c mov bx,[bp+6] ; second parameter to function
5510 \c mov sp,bp ; undo "sub sp,0x40" above
5512 \c retf 4 ; total size of params is 4
5514 At the other end of the process, to call a Pascal function from your
5515 assembly code, you would do something like this:
5519 \c ; and then, further down...
5521 \c push word seg mystring ; Now push the segment, and...
5522 \c push word mystring ; ... offset of "mystring"
5523 \c push word [myint] ; one of my variables
5524 \c call far SomeFunc
5526 This is equivalent to the Pascal code
5528 \c procedure SomeFunc(String: PChar; Int: Integer);
5529 \c SomeFunc(@mystring, myint);
5532 \S{16bpseg} Borland Pascal \I{segment names, Borland Pascal}Segment
5535 Since Borland Pascal's internal unit file format is completely
5536 different from \c{OBJ}, it only makes a very sketchy job of actually
5537 reading and understanding the various information contained in a
5538 real \c{OBJ} file when it links that in. Therefore an object file
5539 intended to be linked to a Pascal program must obey a number of
5542 \b Procedures and functions must be in a segment whose name is
5543 either \c{CODE}, \c{CSEG}, or something ending in \c{_TEXT}.
5545 \b initialized data must be in a segment whose name is either
5546 \c{CONST} or something ending in \c{_DATA}.
5548 \b Uninitialized data must be in a segment whose name is either
5549 \c{DATA}, \c{DSEG}, or something ending in \c{_BSS}.
5551 \b Any other segments in the object file are completely ignored.
5552 \c{GROUP} directives and segment attributes are also ignored.
5555 \S{16bpmacro} Using \i\c{c16.mac} With Pascal Programs
5557 The \c{c16.mac} macro package, described in \k{16cmacro}, can also
5558 be used to simplify writing functions to be called from Pascal
5559 programs, if you code \I\c{PASCAL}\c{%define PASCAL}. This
5560 definition ensures that functions are far (it implies
5561 \i\c{FARCODE}), and also causes procedure return instructions to be
5562 generated with an operand.
5564 Defining \c{PASCAL} does not change the code which calculates the
5565 argument offsets; you must declare your function's arguments in
5566 reverse order. For example:
5574 \c mov ax,[bp + %$i]
5575 \c mov bx,[bp + %$j]
5576 \c mov es,[bp + %$j + 2]
5581 This defines the same routine, conceptually, as the example in
5582 \k{16cmacro}: it defines a function taking two arguments, an integer
5583 and a pointer to an integer, which returns the sum of the integer
5584 and the contents of the pointer. The only difference between this
5585 code and the large-model C version is that \c{PASCAL} is defined
5586 instead of \c{FARCODE}, and that the arguments are declared in
5590 \C{32bit} Writing 32-bit Code (Unix, Win32, DJGPP)
5592 This chapter attempts to cover some of the common issues involved
5593 when writing 32-bit code, to run under \i{Win32} or Unix, or to be
5594 linked with C code generated by a Unix-style C compiler such as
5595 \i{DJGPP}. It covers how to write assembly code to interface with
5596 32-bit C routines, and how to write position-independent code for
5599 Almost all 32-bit code, and in particular all code running under
5600 \c{Win32}, \c{DJGPP} or any of the PC Unix variants, runs in \I{flat
5601 memory model}\e{flat} memory model. This means that the segment registers
5602 and paging have already been set up to give you the same 32-bit 4Gb
5603 address space no matter what segment you work relative to, and that
5604 you should ignore all segment registers completely. When writing
5605 flat-model application code, you never need to use a segment
5606 override or modify any segment register, and the code-section
5607 addresses you pass to \c{CALL} and \c{JMP} live in the same address
5608 space as the data-section addresses you access your variables by and
5609 the stack-section addresses you access local variables and procedure
5610 parameters by. Every address is 32 bits long and contains only an
5614 \H{32c} Interfacing to 32-bit C Programs
5616 A lot of the discussion in \k{16c}, about interfacing to 16-bit C
5617 programs, still applies when working in 32 bits. The absence of
5618 memory models or segmentation worries simplifies things a lot.
5621 \S{32cunder} External Symbol Names
5623 Most 32-bit C compilers share the convention used by 16-bit
5624 compilers, that the names of all global symbols (functions or data)
5625 they define are formed by prefixing an underscore to the name as it
5626 appears in the C program. However, not all of them do: the \c{ELF}
5627 specification states that C symbols do \e{not} have a leading
5628 underscore on their assembly-language names.
5630 The older Linux \c{a.out} C compiler, all \c{Win32} compilers,
5631 \c{DJGPP}, and \c{NetBSD} and \c{FreeBSD}, all use the leading
5632 underscore; for these compilers, the macros \c{cextern} and
5633 \c{cglobal}, as given in \k{16cunder}, will still work. For \c{ELF},
5634 though, the leading underscore should not be used.
5636 See also \k{opt-pfix}.
5638 \S{32cfunc} Function Definitions and Function Calls
5640 \I{functions, C calling convention}The \i{C calling convention}The C
5641 calling convention in 32-bit programs is as follows. In the
5642 following description, the words \e{caller} and \e{callee} are used
5643 to denote the function doing the calling and the function which gets
5646 \b The caller pushes the function's parameters on the stack, one
5647 after another, in reverse order (right to left, so that the first
5648 argument specified to the function is pushed last).
5650 \b The caller then executes a near \c{CALL} instruction to pass
5651 control to the callee.
5653 \b The callee receives control, and typically (although this is not
5654 actually necessary, in functions which do not need to access their
5655 parameters) starts by saving the value of \c{ESP} in \c{EBP} so as
5656 to be able to use \c{EBP} as a base pointer to find its parameters
5657 on the stack. However, the caller was probably doing this too, so
5658 part of the calling convention states that \c{EBP} must be preserved
5659 by any C function. Hence the callee, if it is going to set up
5660 \c{EBP} as a \i{frame pointer}, must push the previous value first.
5662 \b The callee may then access its parameters relative to \c{EBP}.
5663 The doubleword at \c{[EBP]} holds the previous value of \c{EBP} as
5664 it was pushed; the next doubleword, at \c{[EBP+4]}, holds the return
5665 address, pushed implicitly by \c{CALL}. The parameters start after
5666 that, at \c{[EBP+8]}. The leftmost parameter of the function, since
5667 it was pushed last, is accessible at this offset from \c{EBP}; the
5668 others follow, at successively greater offsets. Thus, in a function
5669 such as \c{printf} which takes a variable number of parameters, the
5670 pushing of the parameters in reverse order means that the function
5671 knows where to find its first parameter, which tells it the number
5672 and type of the remaining ones.
5674 \b The callee may also wish to decrease \c{ESP} further, so as to
5675 allocate space on the stack for local variables, which will then be
5676 accessible at negative offsets from \c{EBP}.
5678 \b The callee, if it wishes to return a value to the caller, should
5679 leave the value in \c{AL}, \c{AX} or \c{EAX} depending on the size
5680 of the value. Floating-point results are typically returned in
5683 \b Once the callee has finished processing, it restores \c{ESP} from
5684 \c{EBP} if it had allocated local stack space, then pops the previous
5685 value of \c{EBP}, and returns via \c{RET} (equivalently, \c{RETN}).
5687 \b When the caller regains control from the callee, the function
5688 parameters are still on the stack, so it typically adds an immediate
5689 constant to \c{ESP} to remove them (instead of executing a number of
5690 slow \c{POP} instructions). Thus, if a function is accidentally
5691 called with the wrong number of parameters due to a prototype
5692 mismatch, the stack will still be returned to a sensible state since
5693 the caller, which \e{knows} how many parameters it pushed, does the
5696 There is an alternative calling convention used by Win32 programs
5697 for Windows API calls, and also for functions called \e{by} the
5698 Windows API such as window procedures: they follow what Microsoft
5699 calls the \c{__stdcall} convention. This is slightly closer to the
5700 Pascal convention, in that the callee clears the stack by passing a
5701 parameter to the \c{RET} instruction. However, the parameters are
5702 still pushed in right-to-left order.
5704 Thus, you would define a function in C style in the following way:
5711 \c sub esp,0x40 ; 64 bytes of local stack space
5712 \c mov ebx,[ebp+8] ; first parameter to function
5716 \c leave ; mov esp,ebp / pop ebp
5719 At the other end of the process, to call a C function from your
5720 assembly code, you would do something like this:
5724 \c ; and then, further down...
5726 \c push dword [myint] ; one of my integer variables
5727 \c push dword mystring ; pointer into my data segment
5729 \c add esp,byte 8 ; `byte' saves space
5731 \c ; then those data items...
5736 \c mystring db 'This number -> %d <- should be 1234',10,0
5738 This piece of code is the assembly equivalent of the C code
5740 \c int myint = 1234;
5741 \c printf("This number -> %d <- should be 1234\n", myint);
5744 \S{32cdata} Accessing Data Items
5746 To get at the contents of C variables, or to declare variables which
5747 C can access, you need only declare the names as \c{GLOBAL} or
5748 \c{EXTERN}. (Again, the names require leading underscores, as stated
5749 in \k{32cunder}.) Thus, a C variable declared as \c{int i} can be
5750 accessed from assembler as
5755 And to declare your own integer variable which C programs can access
5756 as \c{extern int j}, you do this (making sure you are assembling in
5757 the \c{_DATA} segment, if necessary):
5762 To access a C array, you need to know the size of the components of
5763 the array. For example, \c{int} variables are four bytes long, so if
5764 a C program declares an array as \c{int a[10]}, you can access
5765 \c{a[3]} by coding \c{mov ax,[_a+12]}. (The byte offset 12 is obtained
5766 by multiplying the desired array index, 3, by the size of the array
5767 element, 4.) The sizes of the C base types in 32-bit compilers are:
5768 1 for \c{char}, 2 for \c{short}, 4 for \c{int}, \c{long} and
5769 \c{float}, and 8 for \c{double}. Pointers, being 32-bit addresses,
5770 are also 4 bytes long.
5772 To access a C \i{data structure}, you need to know the offset from
5773 the base of the structure to the field you are interested in. You
5774 can either do this by converting the C structure definition into a
5775 NASM structure definition (using \c{STRUC}), or by calculating the
5776 one offset and using just that.
5778 To do either of these, you should read your C compiler's manual to
5779 find out how it organizes data structures. NASM gives no special
5780 alignment to structure members in its own \i\c{STRUC} macro, so you
5781 have to specify alignment yourself if the C compiler generates it.
5782 Typically, you might find that a structure like
5789 might be eight bytes long rather than five, since the \c{int} field
5790 would be aligned to a four-byte boundary. However, this sort of
5791 feature is sometimes a configurable option in the C compiler, either
5792 using command-line options or \c{#pragma} lines, so you have to find
5793 out how your own compiler does it.
5796 \S{32cmacro} \i\c{c32.mac}: Helper Macros for the 32-bit C Interface
5798 Included in the NASM archives, in the \I{misc directory}\c{misc}
5799 directory, is a file \c{c32.mac} of macros. It defines three macros:
5800 \i\c{proc}, \i\c{arg} and \i\c{endproc}. These are intended to be
5801 used for C-style procedure definitions, and they automate a lot of
5802 the work involved in keeping track of the calling convention.
5804 An example of an assembly function using the macro set is given
5811 \c mov eax,[ebp + %$i]
5812 \c mov ebx,[ebp + %$j]
5817 This defines \c{_proc32} to be a procedure taking two arguments, the
5818 first (\c{i}) an integer and the second (\c{j}) a pointer to an
5819 integer. It returns \c{i + *j}.
5821 Note that the \c{arg} macro has an \c{EQU} as the first line of its
5822 expansion, and since the label before the macro call gets prepended
5823 to the first line of the expanded macro, the \c{EQU} works, defining
5824 \c{%$i} to be an offset from \c{BP}. A context-local variable is
5825 used, local to the context pushed by the \c{proc} macro and popped
5826 by the \c{endproc} macro, so that the same argument name can be used
5827 in later procedures. Of course, you don't \e{have} to do that.
5829 \c{arg} can take an optional parameter, giving the size of the
5830 argument. If no size is given, 4 is assumed, since it is likely that
5831 many function parameters will be of type \c{int} or pointers.
5834 \H{picdll} Writing NetBSD/FreeBSD/OpenBSD and Linux/ELF \i{Shared
5837 \c{ELF} replaced the older \c{a.out} object file format under Linux
5838 because it contains support for \i{position-independent code}
5839 (\i{PIC}), which makes writing shared libraries much easier. NASM
5840 supports the \c{ELF} position-independent code features, so you can
5841 write Linux \c{ELF} shared libraries in NASM.
5843 \i{NetBSD}, and its close cousins \i{FreeBSD} and \i{OpenBSD}, take
5844 a different approach by hacking PIC support into the \c{a.out}
5845 format. NASM supports this as the \i\c{aoutb} output format, so you
5846 can write \i{BSD} shared libraries in NASM too.
5848 The operating system loads a PIC shared library by memory-mapping
5849 the library file at an arbitrarily chosen point in the address space
5850 of the running process. The contents of the library's code section
5851 must therefore not depend on where it is loaded in memory.
5853 Therefore, you cannot get at your variables by writing code like
5856 \c mov eax,[myvar] ; WRONG
5858 Instead, the linker provides an area of memory called the
5859 \i\e{global offset table}, or \i{GOT}; the GOT is situated at a
5860 constant distance from your library's code, so if you can find out
5861 where your library is loaded (which is typically done using a
5862 \c{CALL} and \c{POP} combination), you can obtain the address of the
5863 GOT, and you can then load the addresses of your variables out of
5864 linker-generated entries in the GOT.
5866 The \e{data} section of a PIC shared library does not have these
5867 restrictions: since the data section is writable, it has to be
5868 copied into memory anyway rather than just paged in from the library
5869 file, so as long as it's being copied it can be relocated too. So
5870 you can put ordinary types of relocation in the data section without
5871 too much worry (but see \k{picglobal} for a caveat).
5874 \S{picgot} Obtaining the Address of the GOT
5876 Each code module in your shared library should define the GOT as an
5879 \c extern _GLOBAL_OFFSET_TABLE_ ; in ELF
5880 \c extern __GLOBAL_OFFSET_TABLE_ ; in BSD a.out
5882 At the beginning of any function in your shared library which plans
5883 to access your data or BSS sections, you must first calculate the
5884 address of the GOT. This is typically done by writing the function
5893 \c add ebx,_GLOBAL_OFFSET_TABLE_+$$-.get_GOT wrt ..gotpc
5895 \c ; the function body comes here
5902 (For BSD, again, the symbol \c{_GLOBAL_OFFSET_TABLE} requires a
5903 second leading underscore.)
5905 The first two lines of this function are simply the standard C
5906 prologue to set up a stack frame, and the last three lines are
5907 standard C function epilogue. The third line, and the fourth to last
5908 line, save and restore the \c{EBX} register, because PIC shared
5909 libraries use this register to store the address of the GOT.
5911 The interesting bit is the \c{CALL} instruction and the following
5912 two lines. The \c{CALL} and \c{POP} combination obtains the address
5913 of the label \c{.get_GOT}, without having to know in advance where
5914 the program was loaded (since the \c{CALL} instruction is encoded
5915 relative to the current position). The \c{ADD} instruction makes use
5916 of one of the special PIC relocation types: \i{GOTPC relocation}.
5917 With the \i\c{WRT ..gotpc} qualifier specified, the symbol
5918 referenced (here \c{_GLOBAL_OFFSET_TABLE_}, the special symbol
5919 assigned to the GOT) is given as an offset from the beginning of the
5920 section. (Actually, \c{ELF} encodes it as the offset from the operand
5921 field of the \c{ADD} instruction, but NASM simplifies this
5922 deliberately, so you do things the same way for both \c{ELF} and
5923 \c{BSD}.) So the instruction then \e{adds} the beginning of the section,
5924 to get the real address of the GOT, and subtracts the value of
5925 \c{.get_GOT} which it knows is in \c{EBX}. Therefore, by the time
5926 that instruction has finished, \c{EBX} contains the address of the GOT.
5928 If you didn't follow that, don't worry: it's never necessary to
5929 obtain the address of the GOT by any other means, so you can put
5930 those three instructions into a macro and safely ignore them:
5937 \c add ebx,_GLOBAL_OFFSET_TABLE_+$$-%%getgot wrt ..gotpc
5941 \S{piclocal} Finding Your Local Data Items
5943 Having got the GOT, you can then use it to obtain the addresses of
5944 your data items. Most variables will reside in the sections you have
5945 declared; they can be accessed using the \I{GOTOFF
5946 relocation}\c{..gotoff} special \I\c{WRT ..gotoff}\c{WRT} type. The
5947 way this works is like this:
5949 \c lea eax,[ebx+myvar wrt ..gotoff]
5951 The expression \c{myvar wrt ..gotoff} is calculated, when the shared
5952 library is linked, to be the offset to the local variable \c{myvar}
5953 from the beginning of the GOT. Therefore, adding it to \c{EBX} as
5954 above will place the real address of \c{myvar} in \c{EAX}.
5956 If you declare variables as \c{GLOBAL} without specifying a size for
5957 them, they are shared between code modules in the library, but do
5958 not get exported from the library to the program that loaded it.
5959 They will still be in your ordinary data and BSS sections, so you
5960 can access them in the same way as local variables, using the above
5961 \c{..gotoff} mechanism.
5963 Note that due to a peculiarity of the way BSD \c{a.out} format
5964 handles this relocation type, there must be at least one non-local
5965 symbol in the same section as the address you're trying to access.
5968 \S{picextern} Finding External and Common Data Items
5970 If your library needs to get at an external variable (external to
5971 the \e{library}, not just to one of the modules within it), you must
5972 use the \I{GOT relocations}\I\c{WRT ..got}\c{..got} type to get at
5973 it. The \c{..got} type, instead of giving you the offset from the
5974 GOT base to the variable, gives you the offset from the GOT base to
5975 a GOT \e{entry} containing the address of the variable. The linker
5976 will set up this GOT entry when it builds the library, and the
5977 dynamic linker will place the correct address in it at load time. So
5978 to obtain the address of an external variable \c{extvar} in \c{EAX},
5981 \c mov eax,[ebx+extvar wrt ..got]
5983 This loads the address of \c{extvar} out of an entry in the GOT. The
5984 linker, when it builds the shared library, collects together every
5985 relocation of type \c{..got}, and builds the GOT so as to ensure it
5986 has every necessary entry present.
5988 Common variables must also be accessed in this way.
5991 \S{picglobal} Exporting Symbols to the Library User
5993 If you want to export symbols to the user of the library, you have
5994 to declare whether they are functions or data, and if they are data,
5995 you have to give the size of the data item. This is because the
5996 dynamic linker has to build \I{PLT}\i{procedure linkage table}
5997 entries for any exported functions, and also moves exported data
5998 items away from the library's data section in which they were
6001 So to export a function to users of the library, you must use
6003 \c global func:function ; declare it as a function
6009 And to export a data item such as an array, you would have to code
6011 \c global array:data array.end-array ; give the size too
6016 Be careful: If you export a variable to the library user, by
6017 declaring it as \c{GLOBAL} and supplying a size, the variable will
6018 end up living in the data section of the main program, rather than
6019 in your library's data section, where you declared it. So you will
6020 have to access your own global variable with the \c{..got} mechanism
6021 rather than \c{..gotoff}, as if it were external (which,
6022 effectively, it has become).
6024 Equally, if you need to store the address of an exported global in
6025 one of your data sections, you can't do it by means of the standard
6028 \c dataptr: dd global_data_item ; WRONG
6030 NASM will interpret this code as an ordinary relocation, in which
6031 \c{global_data_item} is merely an offset from the beginning of the
6032 \c{.data} section (or whatever); so this reference will end up
6033 pointing at your data section instead of at the exported global
6034 which resides elsewhere.
6036 Instead of the above code, then, you must write
6038 \c dataptr: dd global_data_item wrt ..sym
6040 which makes use of the special \c{WRT} type \I\c{WRT ..sym}\c{..sym}
6041 to instruct NASM to search the symbol table for a particular symbol
6042 at that address, rather than just relocating by section base.
6044 Either method will work for functions: referring to one of your
6045 functions by means of
6047 \c funcptr: dd my_function
6049 will give the user the address of the code you wrote, whereas
6051 \c funcptr: dd my_function wrt .sym
6053 will give the address of the procedure linkage table for the
6054 function, which is where the calling program will \e{believe} the
6055 function lives. Either address is a valid way to call the function.
6058 \S{picproc} Calling Procedures Outside the Library
6060 Calling procedures outside your shared library has to be done by
6061 means of a \i\e{procedure linkage table}, or \i{PLT}. The PLT is
6062 placed at a known offset from where the library is loaded, so the
6063 library code can make calls to the PLT in a position-independent
6064 way. Within the PLT there is code to jump to offsets contained in
6065 the GOT, so function calls to other shared libraries or to routines
6066 in the main program can be transparently passed off to their real
6069 To call an external routine, you must use another special PIC
6070 relocation type, \I{PLT relocations}\i\c{WRT ..plt}. This is much
6071 easier than the GOT-based ones: you simply replace calls such as
6072 \c{CALL printf} with the PLT-relative version \c{CALL printf WRT
6076 \S{link} Generating the Library File
6078 Having written some code modules and assembled them to \c{.o} files,
6079 you then generate your shared library with a command such as
6081 \c ld -shared -o library.so module1.o module2.o # for ELF
6082 \c ld -Bshareable -o library.so module1.o module2.o # for BSD
6084 For ELF, if your shared library is going to reside in system
6085 directories such as \c{/usr/lib} or \c{/lib}, it is usually worth
6086 using the \i\c{-soname} flag to the linker, to store the final
6087 library file name, with a version number, into the library:
6089 \c ld -shared -soname library.so.1 -o library.so.1.2 *.o
6091 You would then copy \c{library.so.1.2} into the library directory,
6092 and create \c{library.so.1} as a symbolic link to it.
6095 \C{mixsize} Mixing 16 and 32 Bit Code
6097 This chapter tries to cover some of the issues, largely related to
6098 unusual forms of addressing and jump instructions, encountered when
6099 writing operating system code such as protected-mode initialisation
6100 routines, which require code that operates in mixed segment sizes,
6101 such as code in a 16-bit segment trying to modify data in a 32-bit
6102 one, or jumps between different-size segments.
6105 \H{mixjump} Mixed-Size Jumps\I{jumps, mixed-size}
6107 \I{operating system, writing}\I{writing operating systems}The most
6108 common form of \i{mixed-size instruction} is the one used when
6109 writing a 32-bit OS: having done your setup in 16-bit mode, such as
6110 loading the kernel, you then have to boot it by switching into
6111 protected mode and jumping to the 32-bit kernel start address. In a
6112 fully 32-bit OS, this tends to be the \e{only} mixed-size
6113 instruction you need, since everything before it can be done in pure
6114 16-bit code, and everything after it can be pure 32-bit.
6116 This jump must specify a 48-bit far address, since the target
6117 segment is a 32-bit one. However, it must be assembled in a 16-bit
6118 segment, so just coding, for example,
6120 \c jmp 0x1234:0x56789ABC ; wrong!
6122 will not work, since the offset part of the address will be
6123 truncated to \c{0x9ABC} and the jump will be an ordinary 16-bit far
6126 The Linux kernel setup code gets round the inability of \c{as86} to
6127 generate the required instruction by coding it manually, using
6128 \c{DB} instructions. NASM can go one better than that, by actually
6129 generating the right instruction itself. Here's how to do it right:
6131 \c jmp dword 0x1234:0x56789ABC ; right
6133 \I\c{JMP DWORD}The \c{DWORD} prefix (strictly speaking, it should
6134 come \e{after} the colon, since it is declaring the \e{offset} field
6135 to be a doubleword; but NASM will accept either form, since both are
6136 unambiguous) forces the offset part to be treated as far, in the
6137 assumption that you are deliberately writing a jump from a 16-bit
6138 segment to a 32-bit one.
6140 You can do the reverse operation, jumping from a 32-bit segment to a
6141 16-bit one, by means of the \c{WORD} prefix:
6143 \c jmp word 0x8765:0x4321 ; 32 to 16 bit
6145 If the \c{WORD} prefix is specified in 16-bit mode, or the \c{DWORD}
6146 prefix in 32-bit mode, they will be ignored, since each is
6147 explicitly forcing NASM into a mode it was in anyway.
6150 \H{mixaddr} Addressing Between Different-Size Segments\I{addressing,
6151 mixed-size}\I{mixed-size addressing}
6153 If your OS is mixed 16 and 32-bit, or if you are writing a DOS
6154 extender, you are likely to have to deal with some 16-bit segments
6155 and some 32-bit ones. At some point, you will probably end up
6156 writing code in a 16-bit segment which has to access data in a
6157 32-bit segment, or vice versa.
6159 If the data you are trying to access in a 32-bit segment lies within
6160 the first 64K of the segment, you may be able to get away with using
6161 an ordinary 16-bit addressing operation for the purpose; but sooner
6162 or later, you will want to do 32-bit addressing from 16-bit mode.
6164 The easiest way to do this is to make sure you use a register for
6165 the address, since any effective address containing a 32-bit
6166 register is forced to be a 32-bit address. So you can do
6168 \c mov eax,offset_into_32_bit_segment_specified_by_fs
6169 \c mov dword [fs:eax],0x11223344
6171 This is fine, but slightly cumbersome (since it wastes an
6172 instruction and a register) if you already know the precise offset
6173 you are aiming at. The x86 architecture does allow 32-bit effective
6174 addresses to specify nothing but a 4-byte offset, so why shouldn't
6175 NASM be able to generate the best instruction for the purpose?
6177 It can. As in \k{mixjump}, you need only prefix the address with the
6178 \c{DWORD} keyword, and it will be forced to be a 32-bit address:
6180 \c mov dword [fs:dword my_offset],0x11223344
6182 Also as in \k{mixjump}, NASM is not fussy about whether the
6183 \c{DWORD} prefix comes before or after the segment override, so
6184 arguably a nicer-looking way to code the above instruction is
6186 \c mov dword [dword fs:my_offset],0x11223344
6188 Don't confuse the \c{DWORD} prefix \e{outside} the square brackets,
6189 which controls the size of the data stored at the address, with the
6190 one \c{inside} the square brackets which controls the length of the
6191 address itself. The two can quite easily be different:
6193 \c mov word [dword 0x12345678],0x9ABC
6195 This moves 16 bits of data to an address specified by a 32-bit
6198 You can also specify \c{WORD} or \c{DWORD} prefixes along with the
6199 \c{FAR} prefix to indirect far jumps or calls. For example:
6201 \c call dword far [fs:word 0x4321]
6203 This instruction contains an address specified by a 16-bit offset;
6204 it loads a 48-bit far pointer from that (16-bit segment and 32-bit
6205 offset), and calls that address.
6208 \H{mixother} Other Mixed-Size Instructions
6210 The other way you might want to access data might be using the
6211 string instructions (\c{LODSx}, \c{STOSx} and so on) or the
6212 \c{XLATB} instruction. These instructions, since they take no
6213 parameters, might seem to have no easy way to make them perform
6214 32-bit addressing when assembled in a 16-bit segment.
6216 This is the purpose of NASM's \i\c{a16} and \i\c{a32} prefixes. If
6217 you are coding \c{LODSB} in a 16-bit segment but it is supposed to
6218 be accessing a string in a 32-bit segment, you should load the
6219 desired address into \c{ESI} and then code
6223 The prefix forces the addressing size to 32 bits, meaning that
6224 \c{LODSB} loads from \c{[DS:ESI]} instead of \c{[DS:SI]}. To access
6225 a string in a 16-bit segment when coding in a 32-bit one, the
6226 corresponding \c{a16} prefix can be used.
6228 The \c{a16} and \c{a32} prefixes can be applied to any instruction
6229 in NASM's instruction table, but most of them can generate all the
6230 useful forms without them. The prefixes are necessary only for
6231 instructions with implicit addressing:
6232 \# \c{CMPSx} (\k{insCMPSB}),
6233 \# \c{SCASx} (\k{insSCASB}), \c{LODSx} (\k{insLODSB}), \c{STOSx}
6234 \# (\k{insSTOSB}), \c{MOVSx} (\k{insMOVSB}), \c{INSx} (\k{insINSB}),
6235 \# \c{OUTSx} (\k{insOUTSB}), and \c{XLATB} (\k{insXLATB}).
6236 \c{CMPSx}, \c{SCASx}, \c{LODSx}, \c{STOSx}, \c{MOVSx}, \c{INSx},
6237 \c{OUTSx}, and \c{XLATB}.
6239 various push and pop instructions (\c{PUSHA} and \c{POPF} as well as
6240 the more usual \c{PUSH} and \c{POP}) can accept \c{a16} or \c{a32}
6241 prefixes to force a particular one of \c{SP} or \c{ESP} to be used
6242 as a stack pointer, in case the stack segment in use is a different
6243 size from the code segment.
6245 \c{PUSH} and \c{POP}, when applied to segment registers in 32-bit
6246 mode, also have the slightly odd behaviour that they push and pop 4
6247 bytes at a time, of which the top two are ignored and the bottom two
6248 give the value of the segment register being manipulated. To force
6249 the 16-bit behaviour of segment-register push and pop instructions,
6250 you can use the operand-size prefix \i\c{o16}:
6255 This code saves a doubleword of stack space by fitting two segment
6256 registers into the space which would normally be consumed by pushing
6259 (You can also use the \i\c{o32} prefix to force the 32-bit behaviour
6260 when in 16-bit mode, but this seems less useful.)
6263 \C{64bit} Writing 64-bit Code (Unix, Win64)
6265 This chapter attempts to cover some of the common issues involved when
6266 writing 64-bit code, to run under \i{Win64} or Unix. It covers how to
6267 write assembly code to interface with 64-bit C routines, and how to
6268 write position-independent code for shared libraries.
6270 All 64-bit code uses a flat memory model, since segmentation is not
6271 available in 64-bit mode. The one exception is the \c{FS} and \c{GS}
6272 registers, which still add their bases.
6274 Position independence in 64-bit mode is significantly simpler, since
6275 the processor supports \c{RIP}-relative addressing directly; see the
6276 \c{REL} keyword (\k{effaddr}). On most 64-bit platforms, it is
6277 probably desirable to make that the default, using the directive
6278 \c{DEFAULT REL} (\k{default}).
6280 64-bit programming is relatively similar to 32-bit programming, but
6281 of course pointers are 64 bits long; additionally, all existing
6282 platforms pass arguments in registers rather than on the stack.
6283 Furthermore, 64-bit platforms use SSE2 by default for floating point.
6284 Please see the ABI documentation for your platform.
6286 64-bit platforms differ in the sizes of the fundamental datatypes, not
6287 just from 32-bit platforms but from each other. If a specific size
6288 data type is desired, it is probably best to use the types defined in
6289 the Standard C header \c{<inttypes.h>}.
6291 In 64-bit mode, the default instruction size is still 32 bits. When
6292 loading a value into a 32-bit register (but not an 8- or 16-bit
6293 register), the upper 32 bits of the corresponding 64-bit register are
6296 \H{id64} Immediates and displacements in 64-bit mode
6298 In 64-bit mode, immediates and displacements are generally only 32
6299 bits wide. NASM will therefore truncate most displacements and
6300 immediates to 32 bits.
6302 The only instruction which takes a full \i{64-bit immediate} is:
6306 NASM will produce this instruction whenever the programmer uses
6307 \c{MOV} with an immediate into a 64-bit register. If this is not
6308 desirable, simply specify the equivalent 32-bit register, which will
6309 be automatically zero-extended by the processor, or specify the
6310 immediate as \c{DWORD}:
6312 \c mov rax,foo ; 64-bit immediate
6313 \c mov rax,qword foo ; (identical)
6314 \c mov eax,foo ; 32-bit immediate, zero-extended
6315 \c mov rax,dword foo ; 32-bit immediate, sign-extended
6317 The length of these instructions are 10, 5 and 7 bytes, respectively.
6319 The only instructions which take a full \i{64-bit \e{displacement}} is
6320 loading or storing, using \c{MOV}, \c{AL}, \c{AX}, \c{EAX} or \c{RAX}
6321 (but no other registers) to an absolute 64-bit address. Since this is
6322 a relatively rarely used instruction (64-bit code generally uses
6323 relative addressing), the programmer has to explicitly declare the
6324 displacement size as \c{QWORD}:
6328 \c mov eax,[foo] ; 32-bit absolute disp, sign-extended
6329 \c mov eax,[a32 foo] ; 32-bit absolute disp, zero-extended
6330 \c mov eax,[qword foo] ; 64-bit absolute disp
6334 \c mov eax,[foo] ; 32-bit relative disp
6335 \c mov eax,[a32 foo] ; d:o, address truncated to 32 bits(!)
6336 \c mov eax,[qword foo] ; error
6337 \c mov eax,[abs qword foo] ; 64-bit absolute disp
6339 A sign-extended absolute displacement can access from -2 GB to +2 GB;
6340 a zero-extended absolute displacement can access from 0 to 4 GB.
6342 FIXME: THIS IS NOT YET CORRECTLY IMPLEMENTED
6344 \H{unix64} Interfacing to 64-bit C Programs (Unix)
6346 On Unix, the 64-bit ABI is defined by the document:
6348 \W{http://www.x86-64.org/documentation/abi.pdf}\c{http://www.x86-64.org/documentation/abi.pdf}
6350 Although written for AT&T-syntax assembly, the concepts apply equally
6351 well for NASM-style assembly. What follows is a simplified summary.
6353 The first six integer arguments (from the left) are passed in \c{RDI},
6354 \c{RSI}, \c{RDX}, \c{RCX}, \c{R8}, and \c{R9}, in that order.
6355 Additional integer arguments are passed on the stack. These
6356 registers, plus \c{RAX}, \c{R10} and \c{R11} are destroyed by function
6357 calls, and thus are available for use by the function without saving.
6359 Integer return values are passed in \c{RAX} and \c{RDX}, in that order.
6361 Floating point is done using SSE registers, except for \c{long
6362 double}. Floating-point arguments are passed in \c{XMM0} to \c{XMM7};
6363 return is \c{XMM0} and \c{XMM1}. \c{long double} are passed on the
6364 stack, and returned in \c{ST(0)} and \c{ST(1)}.
6366 All SSE and x87 registers are destroyed by function calls.
6368 On 64-bit Unix, \c{long} is 64 bits.
6370 \H{win64} Interfacing to 64-bit C Programs (Win64)
6372 The Win64 ABI is described at:
6374 \W{http://msdn2.microsoft.com/en-gb/library/ms794533.aspx}\c{http://msdn2.microsoft.com/en-gb/library/ms794533.aspx}
6376 What follows is a simplified summary.
6378 The first four integer arguments are passwd in \c{RCX}, \c{RDX},
6379 \c{R8} and \c{R9}, in that order. Additional integer arguments are
6380 passed on the stack. These registers, plus \c{RAX}, \c{R10} and
6381 \c{R11} are destroyed by function calls, and thus are available for
6382 use by the function without saving.
6384 Integer return values are passed in \c{RAX} only.
6386 Floating point is done using SSE registers, except for \c{long
6387 double}. Floating-point arguments are passed in \c{XMM0} to \c{XMM3};
6388 return is \c{XMM0} only.
6390 On Win64, \c{long} is 32 bits; \c{long long} or \c{_int64} is 64 bits.
6392 \C{trouble} Troubleshooting
6394 This chapter describes some of the common problems that users have
6395 been known to encounter with NASM, and answers them. It also gives
6396 instructions for reporting bugs in NASM if you find a difficulty
6397 that isn't listed here.
6400 \H{problems} Common Problems
6402 \S{inefficient} NASM Generates \i{Inefficient Code}
6404 We sometimes get `bug' reports about NASM generating inefficient, or
6405 even `wrong', code on instructions such as \c{ADD ESP,8}. This is a
6406 deliberate design feature, connected to predictability of output:
6407 NASM, on seeing \c{ADD ESP,8}, will generate the form of the
6408 instruction which leaves room for a 32-bit offset. You need to code
6409 \I\c{BYTE}\c{ADD ESP,BYTE 8} if you want the space-efficient form of
6410 the instruction. This isn't a bug, it's user error: if you prefer to
6411 have NASM produce the more efficient code automatically enable
6412 optimization with the \c{-On} option (see \k{opt-On}).
6415 \S{jmprange} My Jumps are Out of Range\I{out of range, jumps}
6417 Similarly, people complain that when they issue \i{conditional
6418 jumps} (which are \c{SHORT} by default) that try to jump too far,
6419 NASM reports `short jump out of range' instead of making the jumps
6422 This, again, is partly a predictability issue, but in fact has a
6423 more practical reason as well. NASM has no means of being told what
6424 type of processor the code it is generating will be run on; so it
6425 cannot decide for itself that it should generate \i\c{Jcc NEAR} type
6426 instructions, because it doesn't know that it's working for a 386 or
6427 above. Alternatively, it could replace the out-of-range short
6428 \c{JNE} instruction with a very short \c{JE} instruction that jumps
6429 over a \c{JMP NEAR}; this is a sensible solution for processors
6430 below a 386, but hardly efficient on processors which have good
6431 branch prediction \e{and} could have used \c{JNE NEAR} instead. So,
6432 once again, it's up to the user, not the assembler, to decide what
6433 instructions should be generated. See \k{opt-On}.
6436 \S{proborg} \i\c{ORG} Doesn't Work
6438 People writing \i{boot sector} programs in the \c{bin} format often
6439 complain that \c{ORG} doesn't work the way they'd like: in order to
6440 place the \c{0xAA55} signature word at the end of a 512-byte boot
6441 sector, people who are used to MASM tend to code
6445 \c ; some boot sector code
6450 This is not the intended use of the \c{ORG} directive in NASM, and
6451 will not work. The correct way to solve this problem in NASM is to
6452 use the \i\c{TIMES} directive, like this:
6456 \c ; some boot sector code
6458 \c TIMES 510-($-$$) DB 0
6461 The \c{TIMES} directive will insert exactly enough zero bytes into
6462 the output to move the assembly point up to 510. This method also
6463 has the advantage that if you accidentally fill your boot sector too
6464 full, NASM will catch the problem at assembly time and report it, so
6465 you won't end up with a boot sector that you have to disassemble to
6466 find out what's wrong with it.
6469 \S{probtimes} \i\c{TIMES} Doesn't Work
6471 The other common problem with the above code is people who write the
6476 by reasoning that \c{$} should be a pure number, just like 510, so
6477 the difference between them is also a pure number and can happily be
6480 NASM is a \e{modular} assembler: the various component parts are
6481 designed to be easily separable for re-use, so they don't exchange
6482 information unnecessarily. In consequence, the \c{bin} output
6483 format, even though it has been told by the \c{ORG} directive that
6484 the \c{.text} section should start at 0, does not pass that
6485 information back to the expression evaluator. So from the
6486 evaluator's point of view, \c{$} isn't a pure number: it's an offset
6487 from a section base. Therefore the difference between \c{$} and 510
6488 is also not a pure number, but involves a section base. Values
6489 involving section bases cannot be passed as arguments to \c{TIMES}.
6491 The solution, as in the previous section, is to code the \c{TIMES}
6494 \c TIMES 510-($-$$) DB 0
6496 in which \c{$} and \c{$$} are offsets from the same section base,
6497 and so their difference is a pure number. This will solve the
6498 problem and generate sensible code.
6501 \H{bugs} \i{Bugs}\I{reporting bugs}
6503 We have never yet released a version of NASM with any \e{known}
6504 bugs. That doesn't usually stop there being plenty we didn't know
6505 about, though. Any that you find should be reported firstly via the
6507 \W{https://sourceforge.net/projects/nasm/}\c{https://sourceforge.net/projects/nasm/}
6508 (click on "Bugs"), or if that fails then through one of the
6509 contacts in \k{contact}.
6511 Please read \k{qstart} first, and don't report the bug if it's
6512 listed in there as a deliberate feature. (If you think the feature
6513 is badly thought out, feel free to send us reasons why you think it
6514 should be changed, but don't just send us mail saying `This is a
6515 bug' if the documentation says we did it on purpose.) Then read
6516 \k{problems}, and don't bother reporting the bug if it's listed
6519 If you do report a bug, \e{please} give us all of the following
6522 \b What operating system you're running NASM under. DOS, Linux,
6523 NetBSD, Win16, Win32, VMS (I'd be impressed), whatever.
6525 \b If you're running NASM under DOS or Win32, tell us whether you've
6526 compiled your own executable from the DOS source archive, or whether
6527 you were using the standard distribution binaries out of the
6528 archive. If you were using a locally built executable, try to
6529 reproduce the problem using one of the standard binaries, as this
6530 will make it easier for us to reproduce your problem prior to fixing
6533 \b Which version of NASM you're using, and exactly how you invoked
6534 it. Give us the precise command line, and the contents of the
6535 \c{NASMENV} environment variable if any.
6537 \b Which versions of any supplementary programs you're using, and
6538 how you invoked them. If the problem only becomes visible at link
6539 time, tell us what linker you're using, what version of it you've
6540 got, and the exact linker command line. If the problem involves
6541 linking against object files generated by a compiler, tell us what
6542 compiler, what version, and what command line or options you used.
6543 (If you're compiling in an IDE, please try to reproduce the problem
6544 with the command-line version of the compiler.)
6546 \b If at all possible, send us a NASM source file which exhibits the
6547 problem. If this causes copyright problems (e.g. you can only
6548 reproduce the bug in restricted-distribution code) then bear in mind
6549 the following two points: firstly, we guarantee that any source code
6550 sent to us for the purposes of debugging NASM will be used \e{only}
6551 for the purposes of debugging NASM, and that we will delete all our
6552 copies of it as soon as we have found and fixed the bug or bugs in
6553 question; and secondly, we would prefer \e{not} to be mailed large
6554 chunks of code anyway. The smaller the file, the better. A
6555 three-line sample file that does nothing useful \e{except}
6556 demonstrate the problem is much easier to work with than a
6557 fully fledged ten-thousand-line program. (Of course, some errors
6558 \e{do} only crop up in large files, so this may not be possible.)
6560 \b A description of what the problem actually \e{is}. `It doesn't
6561 work' is \e{not} a helpful description! Please describe exactly what
6562 is happening that shouldn't be, or what isn't happening that should.
6563 Examples might be: `NASM generates an error message saying Line 3
6564 for an error that's actually on Line 5'; `NASM generates an error
6565 message that I believe it shouldn't be generating at all'; `NASM
6566 fails to generate an error message that I believe it \e{should} be
6567 generating'; `the object file produced from this source code crashes
6568 my linker'; `the ninth byte of the output file is 66 and I think it
6569 should be 77 instead'.
6571 \b If you believe the output file from NASM to be faulty, send it to
6572 us. That allows us to determine whether our own copy of NASM
6573 generates the same file, or whether the problem is related to
6574 portability issues between our development platforms and yours. We
6575 can handle binary files mailed to us as MIME attachments, uuencoded,
6576 and even BinHex. Alternatively, we may be able to provide an FTP
6577 site you can upload the suspect files to; but mailing them is easier
6580 \b Any other information or data files that might be helpful. If,
6581 for example, the problem involves NASM failing to generate an object
6582 file while TASM can generate an equivalent file without trouble,
6583 then send us \e{both} object files, so we can see what TASM is doing
6584 differently from us.
6587 \A{ndisasm} \i{Ndisasm}
6589 The Netwide Disassembler, NDISASM
6591 \H{ndisintro} Introduction
6594 The Netwide Disassembler is a small companion program to the Netwide
6595 Assembler, NASM. It seemed a shame to have an x86 assembler,
6596 complete with a full instruction table, and not make as much use of
6597 it as possible, so here's a disassembler which shares the
6598 instruction table (and some other bits of code) with NASM.
6600 The Netwide Disassembler does nothing except to produce
6601 disassemblies of \e{binary} source files. NDISASM does not have any
6602 understanding of object file formats, like \c{objdump}, and it will
6603 not understand \c{DOS .EXE} files like \c{debug} will. It just
6607 \H{ndisstart} Getting Started: Installation
6609 See \k{install} for installation instructions. NDISASM, like NASM,
6610 has a \c{man page} which you may want to put somewhere useful, if you
6611 are on a Unix system.
6614 \H{ndisrun} Running NDISASM
6616 To disassemble a file, you will typically use a command of the form
6618 \c ndisasm -b {16|32|64} filename
6620 NDISASM can disassemble 16-, 32- or 64-bit code equally easily,
6621 provided of course that you remember to specify which it is to work
6622 with. If no \i\c{-b} switch is present, NDISASM works in 16-bit mode
6623 by default. The \i\c{-u} switch (for USE32) also invokes 32-bit mode.
6625 Two more command line options are \i\c{-r} which reports the version
6626 number of NDISASM you are running, and \i\c{-h} which gives a short
6627 summary of command line options.
6630 \S{ndiscom} COM Files: Specifying an Origin
6632 To disassemble a \c{DOS .COM} file correctly, a disassembler must assume
6633 that the first instruction in the file is loaded at address \c{0x100},
6634 rather than at zero. NDISASM, which assumes by default that any file
6635 you give it is loaded at zero, will therefore need to be informed of
6638 The \i\c{-o} option allows you to declare a different origin for the
6639 file you are disassembling. Its argument may be expressed in any of
6640 the NASM numeric formats: decimal by default, if it begins with `\c{$}'
6641 or `\c{0x}' or ends in `\c{H}' it's \c{hex}, if it ends in `\c{Q}' it's
6642 \c{octal}, and if it ends in `\c{B}' it's \c{binary}.
6644 Hence, to disassemble a \c{.COM} file:
6646 \c ndisasm -o100h filename.com
6651 \S{ndissync} Code Following Data: Synchronisation
6653 Suppose you are disassembling a file which contains some data which
6654 isn't machine code, and \e{then} contains some machine code. NDISASM
6655 will faithfully plough through the data section, producing machine
6656 instructions wherever it can (although most of them will look
6657 bizarre, and some may have unusual prefixes, e.g. `\c{FS OR AX,0x240A}'),
6658 and generating `DB' instructions ever so often if it's totally stumped.
6659 Then it will reach the code section.
6661 Supposing NDISASM has just finished generating a strange machine
6662 instruction from part of the data section, and its file position is
6663 now one byte \e{before} the beginning of the code section. It's
6664 entirely possible that another spurious instruction will get
6665 generated, starting with the final byte of the data section, and
6666 then the correct first instruction in the code section will not be
6667 seen because the starting point skipped over it. This isn't really
6670 To avoid this, you can specify a `\i\c{synchronisation}' point, or indeed
6671 as many synchronisation points as you like (although NDISASM can
6672 only handle 8192 sync points internally). The definition of a sync
6673 point is this: NDISASM guarantees to hit sync points exactly during
6674 disassembly. If it is thinking about generating an instruction which
6675 would cause it to jump over a sync point, it will discard that
6676 instruction and output a `\c{db}' instead. So it \e{will} start
6677 disassembly exactly from the sync point, and so you \e{will} see all
6678 the instructions in your code section.
6680 Sync points are specified using the \i\c{-s} option: they are measured
6681 in terms of the program origin, not the file position. So if you
6682 want to synchronize after 32 bytes of a \c{.COM} file, you would have to
6685 \c ndisasm -o100h -s120h file.com
6689 \c ndisasm -o100h -s20h file.com
6691 As stated above, you can specify multiple sync markers if you need
6692 to, just by repeating the \c{-s} option.
6695 \S{ndisisync} Mixed Code and Data: Automatic (Intelligent) Synchronisation
6698 Suppose you are disassembling the boot sector of a \c{DOS} floppy (maybe
6699 it has a virus, and you need to understand the virus so that you
6700 know what kinds of damage it might have done you). Typically, this
6701 will contain a \c{JMP} instruction, then some data, then the rest of the
6702 code. So there is a very good chance of NDISASM being \e{misaligned}
6703 when the data ends and the code begins. Hence a sync point is
6706 On the other hand, why should you have to specify the sync point
6707 manually? What you'd do in order to find where the sync point would
6708 be, surely, would be to read the \c{JMP} instruction, and then to use
6709 its target address as a sync point. So can NDISASM do that for you?
6711 The answer, of course, is yes: using either of the synonymous
6712 switches \i\c{-a} (for automatic sync) or \i\c{-i} (for intelligent
6713 sync) will enable \c{auto-sync} mode. Auto-sync mode automatically
6714 generates a sync point for any forward-referring PC-relative jump or
6715 call instruction that NDISASM encounters. (Since NDISASM is one-pass,
6716 if it encounters a PC-relative jump whose target has already been
6717 processed, there isn't much it can do about it...)
6719 Only PC-relative jumps are processed, since an absolute jump is
6720 either through a register (in which case NDISASM doesn't know what
6721 the register contains) or involves a segment address (in which case
6722 the target code isn't in the same segment that NDISASM is working
6723 in, and so the sync point can't be placed anywhere useful).
6725 For some kinds of file, this mechanism will automatically put sync
6726 points in all the right places, and save you from having to place
6727 any sync points manually. However, it should be stressed that
6728 auto-sync mode is \e{not} guaranteed to catch all the sync points, and
6729 you may still have to place some manually.
6731 Auto-sync mode doesn't prevent you from declaring manual sync
6732 points: it just adds automatically generated ones to the ones you
6733 provide. It's perfectly feasible to specify \c{-i} \e{and} some \c{-s}
6736 Another caveat with auto-sync mode is that if, by some unpleasant
6737 fluke, something in your data section should disassemble to a
6738 PC-relative call or jump instruction, NDISASM may obediently place a
6739 sync point in a totally random place, for example in the middle of
6740 one of the instructions in your code section. So you may end up with
6741 a wrong disassembly even if you use auto-sync. Again, there isn't
6742 much I can do about this. If you have problems, you'll have to use
6743 manual sync points, or use the \c{-k} option (documented below) to
6744 suppress disassembly of the data area.
6747 \S{ndisother} Other Options
6749 The \i\c{-e} option skips a header on the file, by ignoring the first N
6750 bytes. This means that the header is \e{not} counted towards the
6751 disassembly offset: if you give \c{-e10 -o10}, disassembly will start
6752 at byte 10 in the file, and this will be given offset 10, not 20.
6754 The \i\c{-k} option is provided with two comma-separated numeric
6755 arguments, the first of which is an assembly offset and the second
6756 is a number of bytes to skip. This \e{will} count the skipped bytes
6757 towards the assembly offset: its use is to suppress disassembly of a
6758 data section which wouldn't contain anything you wanted to see
6762 \H{ndisbugs} Bugs and Improvements
6764 There are no known bugs. However, any you find, with patches if
6765 possible, should be sent to
6766 \W{mailto:nasm-bugs@lists.sourceforge.net}\c{nasm-bugs@lists.sourceforge.net}, or to the
6768 \W{https://sourceforge.net/projects/nasm/}\c{https://sourceforge.net/projects/nasm/}
6769 and we'll try to fix them. Feel free to send contributions and
6770 new features as well.
6772 Future plans include awareness of which processors certain
6773 instructions will run on, and marking of instructions that are too
6774 advanced for some processor (or are \c{FPU} instructions, or are
6775 undocumented opcodes, or are privileged protected-mode instructions,
6780 I hope NDISASM is of some use to somebody. Including me. :-)
6782 I don't recommend taking NDISASM apart to see how an efficient
6783 disassembler works, because as far as I know, it isn't an efficient
6784 one anyway. You have been warned.