1 \input texinfo @c -*-Texinfo-*-
2 @c Copyright (c) 1991, 92, 93, 94, 95, 96, 97, 1998
3 @c Free Software Foundation, Inc.
4 @c UPDATE!! On future updates--
5 @c (1) check for new machine-dep cmdline options in
6 @c md_parse_option definitions in config/tc-*.c
7 @c (2) for platform-specific directives, examine md_pseudo_op
9 @c (3) for object-format specific directives, examine obj_pseudo_op
11 @c (4) portable directives in potable[] in read.c
15 @c defaults, config file may override:
18 @include asconfig.texi
21 @c common OR combinations of conditions
41 @set abnormal-separator
45 @settitle Using @value{AS}
48 @settitle Using @value{AS} (@value{TARGET})
50 @setchapternewpage odd
55 @c WARE! Some of the machine-dependent sections contain tables of machine
56 @c instructions. Except in multi-column format, these tables look silly.
57 @c Unfortunately, Texinfo doesn't have a general-purpose multi-col format, so
58 @c the multi-col format is faked within @example sections.
60 @c Again unfortunately, the natural size that fits on a page, for these tables,
61 @c is different depending on whether or not smallbook is turned on.
62 @c This matters, because of order: text flow switches columns at each page
65 @c The format faked in this source works reasonably well for smallbook,
66 @c not well for the default large-page format. This manual expects that if you
67 @c turn on @smallbook, you will also uncomment the "@set SMALL" to enable the
68 @c tables in question. You can turn on one without the other at your
69 @c discretion, of course.
72 @c the insn tables look just as silly in info files regardless of smallbook,
73 @c might as well show 'em anyways.
79 * As: (as). The GNU assembler.
88 This file documents the GNU Assembler "@value{AS}".
90 Copyright (C) 1991, 92, 93, 94, 95, 96, 97, 1998 Free Software Foundation, Inc.
92 Permission is granted to make and distribute verbatim copies of
93 this manual provided the copyright notice and this permission notice
94 are preserved on all copies.
97 Permission is granted to process this file through Tex and print the
98 results, provided the printed document carries copying permission
99 notice identical to this one except for the removal of this paragraph
100 (this paragraph not being relevant to the printed manual).
103 Permission is granted to copy and distribute modified versions of this manual
104 under the conditions for verbatim copying, provided that the entire resulting
105 derived work is distributed under the terms of a permission notice identical to
108 Permission is granted to copy and distribute translations of this manual
109 into another language, under the above conditions for modified versions.
113 @title Using @value{AS}
114 @subtitle The @sc{gnu} Assembler
116 @subtitle for the @value{TARGET} family
119 @subtitle Version @value{VERSION}
122 The Free Software Foundation Inc. thanks The Nice Computer
123 Company of Australia for loaning Dean Elsner to write the
124 first (Vax) version of @code{as} for Project @sc{gnu}.
125 The proprietors, management and staff of TNCCA thank FSF for
126 distracting the boss while they got some work
129 @author Dean Elsner, Jay Fenlason & friends
133 \hfill {\it Using {\tt @value{AS}}}\par
134 \hfill Edited by Cygnus Support\par
136 %"boxit" macro for figures:
137 %Modified from Knuth's ``boxit'' macro from TeXbook (answer to exercise 21.3)
138 \gdef\boxit#1#2{\vbox{\hrule\hbox{\vrule\kern3pt
139 \vbox{\parindent=0pt\parskip=0pt\hsize=#1\kern3pt\strut\hfil
140 #2\hfil\strut\kern3pt}\kern3pt\vrule}\hrule}}%box with visible outline
141 \gdef\ibox#1#2{\hbox to #1{#2\hfil}\kern8pt}% invisible box
144 @vskip 0pt plus 1filll
145 Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 1998 Free Software Foundation, Inc.
147 Permission is granted to make and distribute verbatim copies of
148 this manual provided the copyright notice and this permission notice
149 are preserved on all copies.
151 Permission is granted to copy and distribute modified versions of this manual
152 under the conditions for verbatim copying, provided that the entire resulting
153 derived work is distributed under the terms of a permission notice identical to
156 Permission is granted to copy and distribute translations of this manual
157 into another language, under the above conditions for modified versions.
162 @top Using @value{AS}
164 This file is a user guide to the @sc{gnu} assembler @code{@value{AS}} version
167 This version of the file describes @code{@value{AS}} configured to generate
168 code for @value{TARGET} architectures.
171 * Overview:: Overview
172 * Invoking:: Command-Line Options
174 * Sections:: Sections and Relocation
176 * Expressions:: Expressions
177 * Pseudo Ops:: Assembler Directives
178 * Machine Dependencies:: Machine Dependent Features
179 * Reporting Bugs:: Reporting Bugs
180 * Acknowledgements:: Who Did What
188 This manual is a user guide to the @sc{gnu} assembler @code{@value{AS}}.
190 This version of the manual describes @code{@value{AS}} configured to generate
191 code for @value{TARGET} architectures.
195 @cindex invocation summary
196 @cindex option summary
197 @cindex summary of options
198 Here is a brief summary of how to invoke @code{@value{AS}}. For details,
199 @pxref{Invoking,,Comand-Line Options}.
201 @c We don't use deffn and friends for the following because they seem
202 @c to be limited to one line for the header.
204 @value{AS} [ -a[cdhlns][=file] ] [ -D ] [ --defsym @var{sym}=@var{val} ]
205 [ -f ] [ --gstabs ] [ --gdwarf2 ] [ --help ] [ -I @var{dir} ] [ -J ] [ -K ] [ -L ]
206 [ --keep-locals ] [ -o @var{objfile} ] [ -R ] [ --statistics ] [ -v ]
207 [ -version ] [ --version ] [ -W ] [ --warn ] [ --fatal-warnings ]
210 @c am29k has no machine-dependent assembler options
213 [ -mbig-endian | -mlittle-endian ]
216 [ -m[arm]1 | -m[arm]2 | -m[arm]250 | -m[arm]3 | -m[arm]6 | -m[arm]60 |
217 -m[arm]600 | -m[arm]610 | -m[arm]620 | -m[arm]7[t][[d]m[i]][fe] | -m[arm]70 |
218 -m[arm]700 | -m[arm]710[c] | -m[arm]7100 | -m[arm]7500 | -m[arm]8 |
219 -m[arm]810 | -m[arm]9 | -m[arm]920 | -m[arm]920t | -m[arm]9tdmi |
220 -mstrongarm | -mstrongarm110 | -mstrongarm1100 ]
221 [ -m[arm]v2 | -m[arm]v2a | -m[arm]v3 | -m[arm]v3m | -m[arm]v4 | -m[arm]v4t |
222 -m[arm]v5 | -[arm]v5t ]
224 [ -mfpa10 | -mfpa11 | -mfpe-old | -mno-fpu ]
226 [ -mapcs-32 | -mapcs-26 | -mapcs-float | -mapcs-reentrant ]
227 [ -mthumb-interwork ]
238 @c Hitachi family chips have no machine-dependent assembler options
241 @c HPPA has no machine-dependent assembler options (yet).
247 @c The order here is important. See c-sparc.texi.
248 [ -Av6 | -Av7 | -Av8 | -Asparclet | -Asparclite
249 -Av8plus | -Av8plusa | -Av9 | -Av9a ]
250 [ -xarch=v8plus | -xarch=v8plusa ] [ -bump ] [ -32 | -64 ]
253 @c Z8000 has no machine-dependent assembler options
256 @c see md_parse_option in tc-i960.c
257 [ -ACA | -ACA_A | -ACB | -ACC | -AKA | -AKB | -AKC | -AMC ]
261 [ -l ] [ -m68000 | -m68010 | -m68020 | ... ]
264 [ -jsri2bsr ] [ -sifilter ] [ -relax ]
267 [ -nocpp ] [ -EL ] [ -EB ] [ -G @var{num} ] [ -mcpu=@var{CPU} ]
268 [ -mips1 ] [ -mips2 ] [ -mips3 ] [ -m4650 ] [ -no-m4650 ]
269 [ --trap ] [ --break ]
270 [ --emulation=@var{name} ]
272 [ -- | @var{files} @dots{} ]
277 Turn on listings, in any of a variety of ways:
281 omit false conditionals
284 omit debugging directives
287 include high-level source
293 include macro expansions
296 omit forms processing
302 set the name of the listing file
305 You may combine these options; for example, use @samp{-aln} for assembly
306 listing without forms processing. The @samp{=file} option, if used, must be
307 the last one. By itself, @samp{-a} defaults to @samp{-ahls}.
310 Ignored. This option is accepted for script compatibility with calls to
313 @item --defsym @var{sym}=@var{value}
314 Define the symbol @var{sym} to be @var{value} before assembling the input file.
315 @var{value} must be an integer constant. As in C, a leading @samp{0x}
316 indicates a hexadecimal value, and a leading @samp{0} indicates an octal value.
319 ``fast''---skip whitespace and comment preprocessing (assume source is
323 Generate stabs debugging information for each assembler line. This
324 may help debugging assembler code, if the debugger can handle it.
327 Generate DWARF2 debugging information for each assembler line. This
328 may help debugging assembler code, if the debugger can handle it.
331 Print a summary of the command line options and exit.
334 Add directory @var{dir} to the search list for @code{.include} directives.
337 Don't warn about signed overflow.
340 @ifclear DIFF-TBL-KLUGE
341 This option is accepted but has no effect on the @value{TARGET} family.
343 @ifset DIFF-TBL-KLUGE
344 Issue warnings when difference tables altered for long displacements.
349 Keep (in the symbol table) local symbols. On traditional a.out systems
350 these start with @samp{L}, but different systems have different local
353 @item -o @var{objfile}
354 Name the object-file output from @code{@value{AS}} @var{objfile}.
357 Fold the data section into the text section.
360 Print the maximum space (in bytes) and total time (in seconds) used by
363 @item --strip-local-absolute
364 Remove local absolute symbols from the outgoing symbol table.
368 Print the @code{as} version.
371 Print the @code{as} version and exit.
375 Suppress warning messages.
377 @item --fatal-warnings
378 Treat warnings as errors.
381 Don't suppress warning messages or treat them as errors.
390 Generate an object file even after errors.
392 @item -- | @var{files} @dots{}
393 Standard input, or source files to assemble.
398 The following options are available when @value{AS} is configured for
403 @cindex ARC endianness
404 @cindex endianness, ARC
405 @cindex big endian output, ARC
407 Generate ``big endian'' format output.
409 @cindex little endian output, ARC
410 @item -mlittle-endian
411 Generate ``little endian'' format output.
417 The following options are available when @value{AS} is configured for the ARM
421 @item -m[arm][1|2|3|6|7|8|9][...]
422 Specify which ARM processor variant is the target.
423 @item -m[arm]v[2|2a|3|3m|4|4t|5|5t]
424 Specify which ARM architecture variant is used by the target.
425 @item -mthumb | -mall
426 Enable or disable Thumb only instruction decoding.
427 @item -mfpa10 | -mfpa11 | -mfpe-old | -mno-fpu
428 Select which Floating Point architcture is the target.
429 @item -mapcs-32 | -mapcs-26 | -mapcs-float | -mapcs-reentrant | -moabi
430 Select which procedure calling convention is in use.
432 Select either big-endian (-EB) or little-endian (-EL) output.
433 @item -mthumb-interwork
434 Specify that the code has been generated with interworking between Thumb and
437 Specify that PIC code has been generated.
442 The following options are available when @value{AS} is configured for
445 @cindex D10V optimization
446 @cindex optimization, D10V
448 Optimize output by parallelizing instructions.
453 The following options are available when @value{AS} is configured for a D30V
456 @cindex D30V optimization
457 @cindex optimization, D30V
459 Optimize output by parallelizing instructions.
463 Warn when nops are generated.
465 @cindex D30V nops after 32-bit multiply
467 Warn when a nop after a 32-bit multiply instruction is generated.
472 The following options are available when @value{AS} is configured for the
473 Intel 80960 processor.
476 @item -ACA | -ACA_A | -ACB | -ACC | -AKA | -AKB | -AKC | -AMC
477 Specify which variant of the 960 architecture is the target.
480 Add code to collect statistics about branches taken.
483 Do not alter compare-and-branch instructions for long displacements;
491 The following options are available when @value{AS} is configured for the
492 Motorola 68000 series.
497 Shorten references to undefined symbols, to one word instead of two.
499 @item -m68000 | -m68008 | -m68010 | -m68020 | -m68030 | -m68040 | -m68060
500 @itemx | -m68302 | -m68331 | -m68332 | -m68333 | -m68340 | -mcpu32 | -m5200
501 Specify what processor in the 68000 family is the target. The default
502 is normally the 68020, but this can be changed at configuration time.
504 @item -m68881 | -m68882 | -mno-68881 | -mno-68882
505 The target machine does (or does not) have a floating-point coprocessor.
506 The default is to assume a coprocessor for 68020, 68030, and cpu32. Although
507 the basic 68000 is not compatible with the 68881, a combination of the
508 two can be specified, since it's possible to do emulation of the
509 coprocessor instructions with the main processor.
511 @item -m68851 | -mno-68851
512 The target machine does (or does not) have a memory-management
513 unit coprocessor. The default is to assume an MMU for 68020 and up.
519 The following options are available when @value{AS} is configured for
520 a picoJava processor.
524 @cindex PJ endianness
525 @cindex endianness, PJ
526 @cindex big endian output, PJ
528 Generate ``big endian'' format output.
530 @cindex little endian output, PJ
532 Generate ``little endian'' format output.
539 The following options are available when @code{@value{AS}} is configured
540 for the SPARC architecture:
543 @item -Av6 | -Av7 | -Av8 | -Asparclet | -Asparclite
544 @itemx -Av8plus | -Av8plusa | -Av9 | -Av9a
545 Explicitly select a variant of the SPARC architecture.
547 @samp{-Av8plus} and @samp{-Av8plusa} select a 32 bit environment.
548 @samp{-Av9} and @samp{-Av9a} select a 64 bit environment.
550 @samp{-Av8plusa} and @samp{-Av9a} enable the SPARC V9 instruction set with
551 UltraSPARC extensions.
553 @item -xarch=v8plus | -xarch=v8plusa
554 For compatibility with the Solaris v9 assembler. These options are
555 equivalent to -Av8plus and -Av8plusa, respectively.
558 Warn when the assembler switches to another architecture.
563 The following options are available when @value{AS} is configured for
568 This option sets the largest size of an object that can be referenced
569 implicitly with the @code{gp} register. It is only accepted for targets that
570 use ECOFF format, such as a DECstation running Ultrix. The default value is 8.
572 @cindex MIPS endianness
573 @cindex endianness, MIPS
574 @cindex big endian output, MIPS
576 Generate ``big endian'' format output.
578 @cindex little endian output, MIPS
580 Generate ``little endian'' format output.
586 Generate code for a particular MIPS Instruction Set Architecture level.
587 @samp{-mips1} corresponds to the @sc{r2000} and @sc{r3000} processors,
588 @samp{-mips2} to the @sc{r6000} processor, and @samp{-mips3} to the @sc{r4000}
593 Generate code for the MIPS @sc{r4650} chip. This tells the assembler to accept
594 the @samp{mad} and @samp{madu} instruction, and to not schedule @samp{nop}
595 instructions around accesses to the @samp{HI} and @samp{LO} registers.
596 @samp{-no-m4650} turns off this option.
598 @item -mcpu=@var{CPU}
599 Generate code for a particular MIPS cpu. This has little effect on the
600 assembler, but it is passed by @code{@value{GCC}}.
603 @item --emulation=@var{name}
604 This option causes @code{@value{AS}} to emulate @code{@value{AS}} configured
605 for some other target, in all respects, including output format (choosing
606 between ELF and ECOFF only), handling of pseudo-opcodes which may generate
607 debugging information or store symbol table information, and default
608 endianness. The available configuration names are: @samp{mipsecoff},
609 @samp{mipself}, @samp{mipslecoff}, @samp{mipsbecoff}, @samp{mipslelf},
610 @samp{mipsbelf}. The first two do not alter the default endianness from that
611 of the primary target for which the assembler was configured; the others change
612 the default to little- or big-endian as indicated by the @samp{b} or @samp{l}
613 in the name. Using @samp{-EB} or @samp{-EL} will override the endianness
614 selection in any case.
616 This option is currently supported only when the primary target
617 @code{@value{AS}} is configured for is a MIPS ELF or ECOFF target.
618 Furthermore, the primary target or others specified with
619 @samp{--enable-targets=@dots{}} at configuration time must include support for
620 the other format, if both are to be available. For example, the Irix 5
621 configuration includes support for both.
623 Eventually, this option will support more configurations, with more
624 fine-grained control over the assembler's behavior, and will be supported for
628 @code{@value{AS}} ignores this option. It is accepted for compatibility with
636 Control how to deal with multiplication overflow and division by zero.
637 @samp{--trap} or @samp{--no-break} (which are synonyms) take a trap exception
638 (and only work for Instruction Set Architecture level 2 and higher);
639 @samp{--break} or @samp{--no-trap} (also synonyms, and the default) take a
645 The following options are available when @value{AS} is configured for
651 Enable or disable the JSRI to BSR transformation. By default this is enabled.
652 The command line option @samp{-nojsri2bsr} can be used to disable it.
656 Enable or disable the silicon filter behaviour. By default this is disabled.
657 The default can be overidden by the @samp{-sifilter} command line option.
660 Alter jump instructions for long displacements.
667 * Manual:: Structure of this Manual
668 * GNU Assembler:: The GNU Assembler
669 * Object Formats:: Object File Formats
670 * Command Line:: Command Line
671 * Input Files:: Input Files
672 * Object:: Output (Object) File
673 * Errors:: Error and Warning Messages
677 @section Structure of this Manual
679 @cindex manual, structure and purpose
680 This manual is intended to describe what you need to know to use
681 @sc{gnu} @code{@value{AS}}. We cover the syntax expected in source files, including
682 notation for symbols, constants, and expressions; the directives that
683 @code{@value{AS}} understands; and of course how to invoke @code{@value{AS}}.
686 We also cover special features in the @value{TARGET}
687 configuration of @code{@value{AS}}, including assembler directives.
690 This manual also describes some of the machine-dependent features of
691 various flavors of the assembler.
694 @cindex machine instructions (not covered)
695 On the other hand, this manual is @emph{not} intended as an introduction
696 to programming in assembly language---let alone programming in general!
697 In a similar vein, we make no attempt to introduce the machine
698 architecture; we do @emph{not} describe the instruction set, standard
699 mnemonics, registers or addressing modes that are standard to a
700 particular architecture.
702 You may want to consult the manufacturer's
703 machine architecture manual for this information.
707 For information on the H8/300 machine instruction set, see @cite{H8/300
708 Series Programming Manual} (Hitachi ADE--602--025). For the H8/300H,
709 see @cite{H8/300H Series Programming Manual} (Hitachi).
712 For information on the H8/500 machine instruction set, see @cite{H8/500
713 Series Programming Manual} (Hitachi M21T001).
716 For information on the Hitachi SH machine instruction set, see
717 @cite{SH-Microcomputer User's Manual} (Hitachi Micro Systems, Inc.).
720 For information on the Z8000 machine instruction set, see @cite{Z8000 CPU Technical Manual}
724 @c I think this is premature---doc@cygnus.com, 17jan1991
726 Throughout this manual, we assume that you are running @dfn{GNU},
727 the portable operating system from the @dfn{Free Software
728 Foundation, Inc.}. This restricts our attention to certain kinds of
729 computer (in particular, the kinds of computers that @sc{gnu} can run on);
730 once this assumption is granted examples and definitions need less
733 @code{@value{AS}} is part of a team of programs that turn a high-level
734 human-readable series of instructions into a low-level
735 computer-readable series of instructions. Different versions of
736 @code{@value{AS}} are used for different kinds of computer.
739 @c There used to be a section "Terminology" here, which defined
740 @c "contents", "byte", "word", and "long". Defining "word" to any
741 @c particular size is confusing when the .word directive may generate 16
742 @c bits on one machine and 32 bits on another; in general, for the user
743 @c version of this manual, none of these terms seem essential to define.
744 @c They were used very little even in the former draft of the manual;
745 @c this draft makes an effort to avoid them (except in names of
749 @section The GNU Assembler
751 @sc{gnu} @code{as} is really a family of assemblers.
753 This manual describes @code{@value{AS}}, a member of that family which is
754 configured for the @value{TARGET} architectures.
756 If you use (or have used) the @sc{gnu} assembler on one architecture, you
757 should find a fairly similar environment when you use it on another
758 architecture. Each version has much in common with the others,
759 including object file formats, most assembler directives (often called
760 @dfn{pseudo-ops}) and assembler syntax.@refill
762 @cindex purpose of @sc{gnu} assembler
763 @code{@value{AS}} is primarily intended to assemble the output of the
764 @sc{gnu} C compiler @code{@value{GCC}} for use by the linker
765 @code{@value{LD}}. Nevertheless, we've tried to make @code{@value{AS}}
766 assemble correctly everything that other assemblers for the same
767 machine would assemble.
769 Any exceptions are documented explicitly (@pxref{Machine Dependencies}).
772 @c This remark should appear in generic version of manual; assumption
773 @c here is that generic version sets M680x0.
774 This doesn't mean @code{@value{AS}} always uses the same syntax as another
775 assembler for the same architecture; for example, we know of several
776 incompatible versions of 680x0 assembly language syntax.
779 Unlike older assemblers, @code{@value{AS}} is designed to assemble a source
780 program in one pass of the source file. This has a subtle impact on the
781 @kbd{.org} directive (@pxref{Org,,@code{.org}}).
784 @section Object File Formats
786 @cindex object file format
787 The @sc{gnu} assembler can be configured to produce several alternative
788 object file formats. For the most part, this does not affect how you
789 write assembly language programs; but directives for debugging symbols
790 are typically different in different file formats. @xref{Symbol
791 Attributes,,Symbol Attributes}.
794 On the @value{TARGET}, @code{@value{AS}} is configured to produce
795 @value{OBJ-NAME} format object files.
797 @c The following should exhaust all configs that set MULTI-OBJ, ideally
799 On the @value{TARGET}, @code{@value{AS}} can be configured to produce either
800 @code{a.out} or COFF format object files.
803 On the @value{TARGET}, @code{@value{AS}} can be configured to produce either
804 @code{b.out} or COFF format object files.
807 On the @value{TARGET}, @code{@value{AS}} can be configured to produce either
808 SOM or ELF format object files.
813 @section Command Line
815 @cindex command line conventions
816 After the program name @code{@value{AS}}, the command line may contain
817 options and file names. Options may appear in any order, and may be
818 before, after, or between file names. The order of file names is
821 @cindex standard input, as input file
823 @file{--} (two hyphens) by itself names the standard input file
824 explicitly, as one of the files for @code{@value{AS}} to assemble.
826 @cindex options, command line
827 Except for @samp{--} any command line argument that begins with a
828 hyphen (@samp{-}) is an option. Each option changes the behavior of
829 @code{@value{AS}}. No option changes the way another option works. An
830 option is a @samp{-} followed by one or more letters; the case of
831 the letter is important. All options are optional.
833 Some options expect exactly one file name to follow them. The file
834 name may either immediately follow the option's letter (compatible
835 with older assemblers) or it may be the next command argument (@sc{gnu}
836 standard). These two command lines are equivalent:
839 @value{AS} -o my-object-file.o mumble.s
840 @value{AS} -omy-object-file.o mumble.s
847 @cindex source program
849 We use the phrase @dfn{source program}, abbreviated @dfn{source}, to
850 describe the program input to one run of @code{@value{AS}}. The program may
851 be in one or more files; how the source is partitioned into files
852 doesn't change the meaning of the source.
854 @c I added "con" prefix to "catenation" just to prove I can overcome my
855 @c APL training... doc@cygnus.com
856 The source program is a concatenation of the text in all the files, in the
859 Each time you run @code{@value{AS}} it assembles exactly one source
860 program. The source program is made up of one or more files.
861 (The standard input is also a file.)
863 You give @code{@value{AS}} a command line that has zero or more input file
864 names. The input files are read (from left file name to right). A
865 command line argument (in any position) that has no special meaning
866 is taken to be an input file name.
868 If you give @code{@value{AS}} no file names it attempts to read one input file
869 from the @code{@value{AS}} standard input, which is normally your terminal. You
870 may have to type @key{ctl-D} to tell @code{@value{AS}} there is no more program
873 Use @samp{--} if you need to explicitly name the standard input file
874 in your command line.
876 If the source is empty, @code{@value{AS}} produces a small, empty object
879 @subheading Filenames and Line-numbers
881 @cindex input file linenumbers
882 @cindex line numbers, in input files
883 There are two ways of locating a line in the input file (or files) and
884 either may be used in reporting error messages. One way refers to a line
885 number in a physical file; the other refers to a line number in a
886 ``logical'' file. @xref{Errors, ,Error and Warning Messages}.
888 @dfn{Physical files} are those files named in the command line given
889 to @code{@value{AS}}.
891 @dfn{Logical files} are simply names declared explicitly by assembler
892 directives; they bear no relation to physical files. Logical file names help
893 error messages reflect the original source file, when @code{@value{AS}} source
894 is itself synthesized from other files. @code{@value{AS}} understands the
895 @samp{#} directives emitted by the @code{@value{GCC}} preprocessor. See also
896 @ref{File,,@code{.file}}.
899 @section Output (Object) File
905 Every time you run @code{@value{AS}} it produces an output file, which is
906 your assembly language program translated into numbers. This file
907 is the object file. Its default name is
915 @code{b.out} when @code{@value{AS}} is configured for the Intel 80960.
917 You can give it another name by using the @code{-o} option. Conventionally,
918 object file names end with @file{.o}. The default name is used for historical
919 reasons: older assemblers were capable of assembling self-contained programs
920 directly into a runnable program. (For some formats, this isn't currently
921 possible, but it can be done for the @code{a.out} format.)
925 The object file is meant for input to the linker @code{@value{LD}}. It contains
926 assembled program code, information to help @code{@value{LD}} integrate
927 the assembled program into a runnable file, and (optionally) symbolic
928 information for the debugger.
930 @c link above to some info file(s) like the description of a.out.
931 @c don't forget to describe @sc{gnu} info as well as Unix lossage.
934 @section Error and Warning Messages
936 @cindex error messsages
937 @cindex warning messages
938 @cindex messages from assembler
939 @code{@value{AS}} may write warnings and error messages to the standard error
940 file (usually your terminal). This should not happen when a compiler
941 runs @code{@value{AS}} automatically. Warnings report an assumption made so
942 that @code{@value{AS}} could keep assembling a flawed program; errors report a
943 grave problem that stops the assembly.
945 @cindex format of warning messages
946 Warning messages have the format
949 file_name:@b{NNN}:Warning Message Text
953 @cindex line numbers, in warnings/errors
954 (where @b{NNN} is a line number). If a logical file name has been given
955 (@pxref{File,,@code{.file}}) it is used for the filename, otherwise the name of
956 the current input file is used. If a logical line number was given
958 (@pxref{Line,,@code{.line}})
962 (@pxref{Line,,@code{.line}})
965 (@pxref{Ln,,@code{.ln}})
968 then it is used to calculate the number printed,
969 otherwise the actual line in the current source file is printed. The
970 message text is intended to be self explanatory (in the grand Unix
973 @cindex format of error messages
974 Error messages have the format
976 file_name:@b{NNN}:FATAL:Error Message Text
978 The file name and line number are derived as for warning
979 messages. The actual message text may be rather less explanatory
980 because many of them aren't supposed to happen.
983 @chapter Command-Line Options
985 @cindex options, all versions of assembler
986 This chapter describes command-line options available in @emph{all}
987 versions of the @sc{gnu} assembler; @pxref{Machine Dependencies}, for options specific
989 to the @value{TARGET}.
992 to particular machine architectures.
995 If you are invoking @code{@value{AS}} via the @sc{gnu} C compiler (version 2),
996 you can use the @samp{-Wa} option to pass arguments through to the assembler.
997 The assembler arguments must be separated from each other (and the @samp{-Wa})
998 by commas. For example:
1001 gcc -c -g -O -Wa,-alh,-L file.c
1005 This passes two options to the assembler: @samp{-alh} (emit a listing to
1006 standard output with with high-level and assembly source) and @samp{-L} (retain
1007 local symbols in the symbol table).
1009 Usually you do not need to use this @samp{-Wa} mechanism, since many compiler
1010 command-line options are automatically passed to the assembler by the compiler.
1011 (You can call the @sc{gnu} compiler driver with the @samp{-v} option to see
1012 precisely what options it passes to each compilation pass, including the
1016 * a:: -a[cdhlns] enable listings
1017 * D:: -D for compatibility
1018 * f:: -f to work faster
1019 * I:: -I for .include search path
1020 @ifclear DIFF-TBL-KLUGE
1021 * K:: -K for compatibility
1023 @ifset DIFF-TBL-KLUGE
1024 * K:: -K for difference tables
1027 * L:: -L to retain local labels
1028 * M:: -M or --mri to assemble in MRI compatibility mode
1029 * MD:: --MD for dependency tracking
1030 * o:: -o to name the object file
1031 * R:: -R to join data and text sections
1032 * statistics:: --statistics to see statistics about assembly
1033 * traditional-format:: --traditional-format for compatible output
1034 * v:: -v to announce version
1035 * W:: -W, --no-warn, --warn, --fatal-warnings to control warnings
1036 * Z:: -Z to make object file even after errors
1040 @section Enable Listings: @code{-a[cdhlns]}
1049 @cindex listings, enabling
1050 @cindex assembly listings, enabling
1052 These options enable listing output from the assembler. By itself,
1053 @samp{-a} requests high-level, assembly, and symbols listing.
1054 You can use other letters to select specific options for the list:
1055 @samp{-ah} requests a high-level language listing,
1056 @samp{-al} requests an output-program assembly listing, and
1057 @samp{-as} requests a symbol table listing.
1058 High-level listings require that a compiler debugging option like
1059 @samp{-g} be used, and that assembly listings (@samp{-al}) be requested
1062 Use the @samp{-ac} option to omit false conditionals from a listing. Any lines
1063 which are not assembled because of a false @code{.if} (or @code{.ifdef}, or any
1064 other conditional), or a true @code{.if} followed by an @code{.else}, will be
1065 omitted from the listing.
1067 Use the @samp{-ad} option to omit debugging directives from the
1070 Once you have specified one of these options, you can further control
1071 listing output and its appearance using the directives @code{.list},
1072 @code{.nolist}, @code{.psize}, @code{.eject}, @code{.title}, and
1074 The @samp{-an} option turns off all forms processing.
1075 If you do not request listing output with one of the @samp{-a} options, the
1076 listing-control directives have no effect.
1078 The letters after @samp{-a} may be combined into one option,
1079 @emph{e.g.}, @samp{-aln}.
1085 This option has no effect whatsoever, but it is accepted to make it more
1086 likely that scripts written for other assemblers also work with
1090 @section Work Faster: @code{-f}
1093 @cindex trusted compiler
1094 @cindex faster processing (@code{-f})
1095 @samp{-f} should only be used when assembling programs written by a
1096 (trusted) compiler. @samp{-f} stops the assembler from doing whitespace
1097 and comment preprocessing on
1098 the input file(s) before assembling them. @xref{Preprocessing,
1102 @emph{Warning:} if you use @samp{-f} when the files actually need to be
1103 preprocessed (if they contain comments, for example), @code{@value{AS}} does
1108 @section @code{.include} search path: @code{-I} @var{path}
1110 @kindex -I @var{path}
1111 @cindex paths for @code{.include}
1112 @cindex search path for @code{.include}
1113 @cindex @code{include} directive search path
1114 Use this option to add a @var{path} to the list of directories
1115 @code{@value{AS}} searches for files specified in @code{.include}
1116 directives (@pxref{Include,,@code{.include}}). You may use @code{-I} as
1117 many times as necessary to include a variety of paths. The current
1118 working directory is always searched first; after that, @code{@value{AS}}
1119 searches any @samp{-I} directories in the same order as they were
1120 specified (left to right) on the command line.
1123 @section Difference Tables: @code{-K}
1126 @ifclear DIFF-TBL-KLUGE
1127 On the @value{TARGET} family, this option is allowed, but has no effect. It is
1128 permitted for compatibility with the @sc{gnu} assembler on other platforms,
1129 where it can be used to warn when the assembler alters the machine code
1130 generated for @samp{.word} directives in difference tables. The @value{TARGET}
1131 family does not have the addressing limitations that sometimes lead to this
1132 alteration on other platforms.
1135 @ifset DIFF-TBL-KLUGE
1136 @cindex difference tables, warning
1137 @cindex warning for altered difference tables
1138 @code{@value{AS}} sometimes alters the code emitted for directives of the form
1139 @samp{.word @var{sym1}-@var{sym2}}; @pxref{Word,,@code{.word}}.
1140 You can use the @samp{-K} option if you want a warning issued when this
1145 @section Include Local Labels: @code{-L}
1148 @cindex local labels, retaining in output
1149 Labels beginning with @samp{L} (upper case only) are called @dfn{local
1150 labels}. @xref{Symbol Names}. Normally you do not see such labels when
1151 debugging, because they are intended for the use of programs (like
1152 compilers) that compose assembler programs, not for your notice.
1153 Normally both @code{@value{AS}} and @code{@value{LD}} discard such labels, so you do not
1154 normally debug with them.
1156 This option tells @code{@value{AS}} to retain those @samp{L@dots{}} symbols
1157 in the object file. Usually if you do this you also tell the linker
1158 @code{@value{LD}} to preserve symbols whose names begin with @samp{L}.
1160 By default, a local label is any label beginning with @samp{L}, but each
1161 target is allowed to redefine the local label prefix.
1163 On the HPPA local labels begin with @samp{L$}.
1166 @samp{;} for the ARM family;
1170 @section Assemble in MRI Compatibility Mode: @code{-M}
1173 @cindex MRI compatibility mode
1174 The @code{-M} or @code{--mri} option selects MRI compatibility mode. This
1175 changes the syntax and pseudo-op handling of @code{@value{AS}} to make it
1176 compatible with the @code{ASM68K} or the @code{ASM960} (depending upon the
1177 configured target) assembler from Microtec Research. The exact nature of the
1178 MRI syntax will not be documented here; see the MRI manuals for more
1179 information. Note in particular that the handling of macros and macro
1180 arguments is somewhat different. The purpose of this option is to permit
1181 assembling existing MRI assembler code using @code{@value{AS}}.
1183 The MRI compatibility is not complete. Certain operations of the MRI assembler
1184 depend upon its object file format, and can not be supported using other object
1185 file formats. Supporting these would require enhancing each object file format
1186 individually. These are:
1189 @item global symbols in common section
1191 The m68k MRI assembler supports common sections which are merged by the linker.
1192 Other object file formats do not support this. @code{@value{AS}} handles
1193 common sections by treating them as a single common symbol. It permits local
1194 symbols to be defined within a common section, but it can not support global
1195 symbols, since it has no way to describe them.
1197 @item complex relocations
1199 The MRI assemblers support relocations against a negated section address, and
1200 relocations which combine the start addresses of two or more sections. These
1201 are not support by other object file formats.
1203 @item @code{END} pseudo-op specifying start address
1205 The MRI @code{END} pseudo-op permits the specification of a start address.
1206 This is not supported by other object file formats. The start address may
1207 instead be specified using the @code{-e} option to the linker, or in a linker
1210 @item @code{IDNT}, @code{.ident} and @code{NAME} pseudo-ops
1212 The MRI @code{IDNT}, @code{.ident} and @code{NAME} pseudo-ops assign a module
1213 name to the output file. This is not supported by other object file formats.
1215 @item @code{ORG} pseudo-op
1217 The m68k MRI @code{ORG} pseudo-op begins an absolute section at a given
1218 address. This differs from the usual @code{@value{AS}} @code{.org} pseudo-op,
1219 which changes the location within the current section. Absolute sections are
1220 not supported by other object file formats. The address of a section may be
1221 assigned within a linker script.
1224 There are some other features of the MRI assembler which are not supported by
1225 @code{@value{AS}}, typically either because they are difficult or because they
1226 seem of little consequence. Some of these may be supported in future releases.
1230 @item EBCDIC strings
1232 EBCDIC strings are not supported.
1234 @item packed binary coded decimal
1236 Packed binary coded decimal is not supported. This means that the @code{DC.P}
1237 and @code{DCB.P} pseudo-ops are not supported.
1239 @item @code{FEQU} pseudo-op
1241 The m68k @code{FEQU} pseudo-op is not supported.
1243 @item @code{NOOBJ} pseudo-op
1245 The m68k @code{NOOBJ} pseudo-op is not supported.
1247 @item @code{OPT} branch control options
1249 The m68k @code{OPT} branch control options---@code{B}, @code{BRS}, @code{BRB},
1250 @code{BRL}, and @code{BRW}---are ignored. @code{@value{AS}} automatically
1251 relaxes all branches, whether forward or backward, to an appropriate size, so
1252 these options serve no purpose.
1254 @item @code{OPT} list control options
1256 The following m68k @code{OPT} list control options are ignored: @code{C},
1257 @code{CEX}, @code{CL}, @code{CRE}, @code{E}, @code{G}, @code{I}, @code{M},
1258 @code{MEX}, @code{MC}, @code{MD}, @code{X}.
1260 @item other @code{OPT} options
1262 The following m68k @code{OPT} options are ignored: @code{NEST}, @code{O},
1263 @code{OLD}, @code{OP}, @code{P}, @code{PCO}, @code{PCR}, @code{PCS}, @code{R}.
1265 @item @code{OPT} @code{D} option is default
1267 The m68k @code{OPT} @code{D} option is the default, unlike the MRI assembler.
1268 @code{OPT NOD} may be used to turn it off.
1270 @item @code{XREF} pseudo-op.
1272 The m68k @code{XREF} pseudo-op is ignored.
1274 @item @code{.debug} pseudo-op
1276 The i960 @code{.debug} pseudo-op is not supported.
1278 @item @code{.extended} pseudo-op
1280 The i960 @code{.extended} pseudo-op is not supported.
1282 @item @code{.list} pseudo-op.
1284 The various options of the i960 @code{.list} pseudo-op are not supported.
1286 @item @code{.optimize} pseudo-op
1288 The i960 @code{.optimize} pseudo-op is not supported.
1290 @item @code{.output} pseudo-op
1292 The i960 @code{.output} pseudo-op is not supported.
1294 @item @code{.setreal} pseudo-op
1296 The i960 @code{.setreal} pseudo-op is not supported.
1301 @section Dependency tracking: @code{--MD}
1304 @cindex dependency tracking
1307 @code{@value{AS}} can generate a dependency file for the file it creates. This
1308 file consists of a single rule suitable for @code{make} describing the
1309 dependencies of the main source file.
1311 The rule is written to the file named in its argument.
1313 This feature is used in the automatic updating of makefiles.
1316 @section Name the Object File: @code{-o}
1319 @cindex naming object file
1320 @cindex object file name
1321 There is always one object file output when you run @code{@value{AS}}. By
1322 default it has the name
1325 @file{a.out} (or @file{b.out}, for Intel 960 targets only).
1339 You use this option (which takes exactly one filename) to give the
1340 object file a different name.
1342 Whatever the object file is called, @code{@value{AS}} overwrites any
1343 existing file of the same name.
1346 @section Join Data and Text Sections: @code{-R}
1349 @cindex data and text sections, joining
1350 @cindex text and data sections, joining
1351 @cindex joining text and data sections
1352 @cindex merging text and data sections
1353 @code{-R} tells @code{@value{AS}} to write the object file as if all
1354 data-section data lives in the text section. This is only done at
1355 the very last moment: your binary data are the same, but data
1356 section parts are relocated differently. The data section part of
1357 your object file is zero bytes long because all its bytes are
1358 appended to the text section. (@xref{Sections,,Sections and Relocation}.)
1360 When you specify @code{-R} it would be possible to generate shorter
1361 address displacements (because we do not have to cross between text and
1362 data section). We refrain from doing this simply for compatibility with
1363 older versions of @code{@value{AS}}. In future, @code{-R} may work this way.
1366 When @code{@value{AS}} is configured for COFF output,
1367 this option is only useful if you use sections named @samp{.text} and
1372 @code{-R} is not supported for any of the HPPA targets. Using
1373 @code{-R} generates a warning from @code{@value{AS}}.
1377 @section Display Assembly Statistics: @code{--statistics}
1379 @kindex --statistics
1380 @cindex statistics, about assembly
1381 @cindex time, total for assembly
1382 @cindex space used, maximum for assembly
1383 Use @samp{--statistics} to display two statistics about the resources used by
1384 @code{@value{AS}}: the maximum amount of space allocated during the assembly
1385 (in bytes), and the total execution time taken for the assembly (in @sc{cpu}
1388 @node traditional-format
1389 @section Compatible output: @code{--traditional-format}
1391 @kindex --traditional-format
1392 For some targets, the output of @code{@value{AS}} is different in some ways
1393 from the output of some existing assembler. This switch requests
1394 @code{@value{AS}} to use the traditional format instead.
1396 For example, it disables the exception frame optimizations which
1397 @code{@value{AS}} normally does by default on @code{@value{GCC}} output.
1400 @section Announce Version: @code{-v}
1404 @cindex assembler version
1405 @cindex version of assembler
1406 You can find out what version of as is running by including the
1407 option @samp{-v} (which you can also spell as @samp{-version}) on the
1411 @section Control Warnings: @code{-W}, @code{--warn}, @code{--no-warn}, @code{--fatal-warnings}
1413 @code{@value{AS}} should never give a warning or error message when
1414 assembling compiler output. But programs written by people often
1415 cause @code{@value{AS}} to give a warning that a particular assumption was
1416 made. All such warnings are directed to the standard error file.
1419 @kindex @samp{--no-warn}
1420 @cindex suppressing warnings
1421 @cindex warnings, suppressing
1422 If you use the @code{-W} and @code{--no-warn} options, no warnings are issued.
1423 This only affects the warning messages: it does not change any particular of
1424 how @code{@value{AS}} assembles your file. Errors, which stop the assembly,
1427 @kindex @samp{--fatal-warnings}
1428 @cindex errors, caused by warnings
1429 @cindex warnings, causing error
1430 If you use the @code{--fatal-warnings} option, @code{@value{AS}} considers
1431 files that generate warnings to be in error.
1433 @kindex @samp{--warn}
1434 @cindex warnings, switching on
1435 You can switch these options off again by specifying @code{--warn}, which
1436 causes warnings to be output as usual.
1439 @section Generate Object File in Spite of Errors: @code{-Z}
1440 @cindex object file, after errors
1441 @cindex errors, continuing after
1442 After an error message, @code{@value{AS}} normally produces no output. If for
1443 some reason you are interested in object file output even after
1444 @code{@value{AS}} gives an error message on your program, use the @samp{-Z}
1445 option. If there are any errors, @code{@value{AS}} continues anyways, and
1446 writes an object file after a final warning message of the form @samp{@var{n}
1447 errors, @var{m} warnings, generating bad object file.}
1452 @cindex machine-independent syntax
1453 @cindex syntax, machine-independent
1454 This chapter describes the machine-independent syntax allowed in a
1455 source file. @code{@value{AS}} syntax is similar to what many other
1456 assemblers use; it is inspired by the BSD 4.2
1461 assembler, except that @code{@value{AS}} does not assemble Vax bit-fields.
1465 * Preprocessing:: Preprocessing
1466 * Whitespace:: Whitespace
1467 * Comments:: Comments
1468 * Symbol Intro:: Symbols
1469 * Statements:: Statements
1470 * Constants:: Constants
1474 @section Preprocessing
1476 @cindex preprocessing
1477 The @code{@value{AS}} internal preprocessor:
1479 @cindex whitespace, removed by preprocessor
1481 adjusts and removes extra whitespace. It leaves one space or tab before
1482 the keywords on a line, and turns any other whitespace on the line into
1485 @cindex comments, removed by preprocessor
1487 removes all comments, replacing them with a single space, or an
1488 appropriate number of newlines.
1490 @cindex constants, converted by preprocessor
1492 converts character constants into the appropriate numeric values.
1495 It does not do macro processing, include file handling, or
1496 anything else you may get from your C compiler's preprocessor. You can
1497 do include file processing with the @code{.include} directive
1498 (@pxref{Include,,@code{.include}}). You can use the @sc{gnu} C compiler driver
1499 to get other ``CPP'' style preprocessing, by giving the input file a
1500 @samp{.S} suffix. @xref{Overall Options,, Options Controlling the Kind of
1501 Output, gcc.info, Using GNU CC}.
1503 Excess whitespace, comments, and character constants
1504 cannot be used in the portions of the input text that are not
1507 @cindex turning preprocessing on and off
1508 @cindex preprocessing, turning on and off
1511 If the first line of an input file is @code{#NO_APP} or if you use the
1512 @samp{-f} option, whitespace and comments are not removed from the input file.
1513 Within an input file, you can ask for whitespace and comment removal in
1514 specific portions of the by putting a line that says @code{#APP} before the
1515 text that may contain whitespace or comments, and putting a line that says
1516 @code{#NO_APP} after this text. This feature is mainly intend to support
1517 @code{asm} statements in compilers whose output is otherwise free of comments
1524 @dfn{Whitespace} is one or more blanks or tabs, in any order.
1525 Whitespace is used to separate symbols, and to make programs neater for
1526 people to read. Unless within character constants
1527 (@pxref{Characters,,Character Constants}), any whitespace means the same
1528 as exactly one space.
1534 There are two ways of rendering comments to @code{@value{AS}}. In both
1535 cases the comment is equivalent to one space.
1537 Anything from @samp{/*} through the next @samp{*/} is a comment.
1538 This means you may not nest these comments.
1542 The only way to include a newline ('\n') in a comment
1543 is to use this sort of comment.
1546 /* This sort of comment does not nest. */
1549 @cindex line comment character
1550 Anything from the @dfn{line comment} character to the next newline
1551 is considered a comment and is ignored. The line comment character is
1553 @samp{;} for the AMD 29K family;
1556 @samp{;} on the ARC;
1559 @samp{;} for the H8/300 family;
1562 @samp{!} for the H8/500 family;
1565 @samp{;} for the HPPA;
1568 @samp{#} on the i960;
1571 @samp{;} for picoJava;
1574 @samp{!} for the Hitachi SH;
1577 @samp{!} on the SPARC;
1580 @samp{#} on the m32r;
1583 @samp{|} on the 680x0;
1586 @samp{#} on the Vax;
1589 @samp{!} for the Z8000;
1592 @samp{#} on the V850;
1594 see @ref{Machine Dependencies}. @refill
1595 @c FIXME What about i386, m88k, i860?
1598 On some machines there are two different line comment characters. One
1599 character only begins a comment if it is the first non-whitespace character on
1600 a line, while the other always begins a comment.
1604 The V850 assembler also supports a double dash as starting a comment that
1605 extends to the end of the line.
1611 @cindex lines starting with @code{#}
1612 @cindex logical line numbers
1613 To be compatible with past assemblers, lines that begin with @samp{#} have a
1614 special interpretation. Following the @samp{#} should be an absolute
1615 expression (@pxref{Expressions}): the logical line number of the @emph{next}
1616 line. Then a string (@pxref{Strings,, Strings}) is allowed: if present it is a
1617 new logical file name. The rest of the line, if any, should be whitespace.
1619 If the first non-whitespace characters on the line are not numeric,
1620 the line is ignored. (Just like a comment.)
1623 # This is an ordinary comment.
1624 # 42-6 "new_file_name" # New logical file name
1625 # This is logical line # 36.
1627 This feature is deprecated, and may disappear from future versions
1628 of @code{@value{AS}}.
1633 @cindex characters used in symbols
1634 @ifclear SPECIAL-SYMS
1635 A @dfn{symbol} is one or more characters chosen from the set of all
1636 letters (both upper and lower case), digits and the three characters
1642 A @dfn{symbol} is one or more characters chosen from the set of all
1643 letters (both upper and lower case), digits and the three characters
1644 @samp{._$}. (Save that, on the H8/300 only, you may not use @samp{$} in
1650 On most machines, you can also use @code{$} in symbol names; exceptions
1651 are noted in @ref{Machine Dependencies}.
1653 No symbol may begin with a digit. Case is significant.
1654 There is no length limit: all characters are significant. Symbols are
1655 delimited by characters not in that set, or by the beginning of a file
1656 (since the source program must end with a newline, the end of a file is
1657 not a possible symbol delimiter). @xref{Symbols}.
1658 @cindex length of symbols
1663 @cindex statements, structure of
1664 @cindex line separator character
1665 @cindex statement separator character
1667 @ifclear abnormal-separator
1668 A @dfn{statement} ends at a newline character (@samp{\n}) or at a
1669 semicolon (@samp{;}). The newline or semicolon is considered part of
1670 the preceding statement. Newlines and semicolons within character
1671 constants are an exception: they do not end statements.
1673 @ifset abnormal-separator
1675 A @dfn{statement} ends at a newline character (@samp{\n}) or an ``at''
1676 sign (@samp{@@}). The newline or at sign is considered part of the
1677 preceding statement. Newlines and at signs within character constants
1678 are an exception: they do not end statements.
1681 A @dfn{statement} ends at a newline character (@samp{\n}) or an exclamation
1682 point (@samp{!}). The newline or exclamation point is considered part of the
1683 preceding statement. Newlines and exclamation points within character
1684 constants are an exception: they do not end statements.
1687 A @dfn{statement} ends at a newline character (@samp{\n}); or (for the
1688 H8/300) a dollar sign (@samp{$}); or (for the
1691 (@samp{;}). The newline or separator character is considered part of
1692 the preceding statement. Newlines and separators within character
1693 constants are an exception: they do not end statements.
1698 A @dfn{statement} ends at a newline character (@samp{\n}) or line
1699 separator character. (The line separator is usually @samp{;}, unless
1700 this conflicts with the comment character; @pxref{Machine Dependencies}.) The
1701 newline or separator character is considered part of the preceding
1702 statement. Newlines and separators within character constants are an
1703 exception: they do not end statements.
1706 @cindex newline, required at file end
1707 @cindex EOF, newline must precede
1708 It is an error to end any statement with end-of-file: the last
1709 character of any input file should be a newline.@refill
1711 An empty statement is allowed, and may include whitespace. It is ignored.
1713 @cindex instructions and directives
1714 @cindex directives and instructions
1715 @c "key symbol" is not used elsewhere in the document; seems pedantic to
1716 @c @defn{} it in that case, as was done previously... doc@cygnus.com,
1718 A statement begins with zero or more labels, optionally followed by a
1719 key symbol which determines what kind of statement it is. The key
1720 symbol determines the syntax of the rest of the statement. If the
1721 symbol begins with a dot @samp{.} then the statement is an assembler
1722 directive: typically valid for any computer. If the symbol begins with
1723 a letter the statement is an assembly language @dfn{instruction}: it
1724 assembles into a machine language instruction.
1726 Different versions of @code{@value{AS}} for different computers
1727 recognize different instructions. In fact, the same symbol may
1728 represent a different instruction in a different computer's assembly
1732 @cindex @code{:} (label)
1733 @cindex label (@code{:})
1734 A label is a symbol immediately followed by a colon (@code{:}).
1735 Whitespace before a label or after a colon is permitted, but you may not
1736 have whitespace between a label's symbol and its colon. @xref{Labels}.
1739 For HPPA targets, labels need not be immediately followed by a colon, but
1740 the definition of a label must begin in column zero. This also implies that
1741 only one label may be defined on each line.
1745 label: .directive followed by something
1746 another_label: # This is an empty statement.
1747 instruction operand_1, operand_2, @dots{}
1754 A constant is a number, written so that its value is known by
1755 inspection, without knowing any context. Like this:
1758 .byte 74, 0112, 092, 0x4A, 0X4a, 'J, '\J # All the same value.
1759 .ascii "Ring the bell\7" # A string constant.
1760 .octa 0x123456789abcdef0123456789ABCDEF0 # A bignum.
1761 .float 0f-314159265358979323846264338327\
1762 95028841971.693993751E-40 # - pi, a flonum.
1767 * Characters:: Character Constants
1768 * Numbers:: Number Constants
1772 @subsection Character Constants
1774 @cindex character constants
1775 @cindex constants, character
1776 There are two kinds of character constants. A @dfn{character} stands
1777 for one character in one byte and its value may be used in
1778 numeric expressions. String constants (properly called string
1779 @emph{literals}) are potentially many bytes and their values may not be
1780 used in arithmetic expressions.
1784 * Chars:: Characters
1788 @subsubsection Strings
1790 @cindex string constants
1791 @cindex constants, string
1792 A @dfn{string} is written between double-quotes. It may contain
1793 double-quotes or null characters. The way to get special characters
1794 into a string is to @dfn{escape} these characters: precede them with
1795 a backslash @samp{\} character. For example @samp{\\} represents
1796 one backslash: the first @code{\} is an escape which tells
1797 @code{@value{AS}} to interpret the second character literally as a backslash
1798 (which prevents @code{@value{AS}} from recognizing the second @code{\} as an
1799 escape character). The complete list of escapes follows.
1801 @cindex escape codes, character
1802 @cindex character escape codes
1805 @c Mnemonic for ACKnowledge; for ASCII this is octal code 007.
1807 @cindex @code{\b} (backspace character)
1808 @cindex backspace (@code{\b})
1810 Mnemonic for backspace; for ASCII this is octal code 010.
1813 @c Mnemonic for EOText; for ASCII this is octal code 004.
1815 @cindex @code{\f} (formfeed character)
1816 @cindex formfeed (@code{\f})
1818 Mnemonic for FormFeed; for ASCII this is octal code 014.
1820 @cindex @code{\n} (newline character)
1821 @cindex newline (@code{\n})
1823 Mnemonic for newline; for ASCII this is octal code 012.
1826 @c Mnemonic for prefix; for ASCII this is octal code 033, usually known as @code{escape}.
1828 @cindex @code{\r} (carriage return character)
1829 @cindex carriage return (@code{\r})
1831 Mnemonic for carriage-Return; for ASCII this is octal code 015.
1834 @c Mnemonic for space; for ASCII this is octal code 040. Included for compliance with
1835 @c other assemblers.
1837 @cindex @code{\t} (tab)
1838 @cindex tab (@code{\t})
1840 Mnemonic for horizontal Tab; for ASCII this is octal code 011.
1843 @c Mnemonic for Vertical tab; for ASCII this is octal code 013.
1844 @c @item \x @var{digit} @var{digit} @var{digit}
1845 @c A hexadecimal character code. The numeric code is 3 hexadecimal digits.
1847 @cindex @code{\@var{ddd}} (octal character code)
1848 @cindex octal character code (@code{\@var{ddd}})
1849 @item \ @var{digit} @var{digit} @var{digit}
1850 An octal character code. The numeric code is 3 octal digits.
1851 For compatibility with other Unix systems, 8 and 9 are accepted as digits:
1852 for example, @code{\008} has the value 010, and @code{\009} the value 011.
1854 @cindex @code{\@var{xd...}} (hex character code)
1855 @cindex hex character code (@code{\@var{xd...}})
1856 @item \@code{x} @var{hex-digits...}
1857 A hex character code. All trailing hex digits are combined. Either upper or
1858 lower case @code{x} works.
1860 @cindex @code{\\} (@samp{\} character)
1861 @cindex backslash (@code{\\})
1863 Represents one @samp{\} character.
1866 @c Represents one @samp{'} (accent acute) character.
1867 @c This is needed in single character literals
1868 @c (@xref{Characters,,Character Constants}.) to represent
1871 @cindex @code{\"} (doublequote character)
1872 @cindex doublequote (@code{\"})
1874 Represents one @samp{"} character. Needed in strings to represent
1875 this character, because an unescaped @samp{"} would end the string.
1877 @item \ @var{anything-else}
1878 Any other character when escaped by @kbd{\} gives a warning, but
1879 assembles as if the @samp{\} was not present. The idea is that if
1880 you used an escape sequence you clearly didn't want the literal
1881 interpretation of the following character. However @code{@value{AS}} has no
1882 other interpretation, so @code{@value{AS}} knows it is giving you the wrong
1883 code and warns you of the fact.
1886 Which characters are escapable, and what those escapes represent,
1887 varies widely among assemblers. The current set is what we think
1888 the BSD 4.2 assembler recognizes, and is a subset of what most C
1889 compilers recognize. If you are in doubt, do not use an escape
1893 @subsubsection Characters
1895 @cindex single character constant
1896 @cindex character, single
1897 @cindex constant, single character
1898 A single character may be written as a single quote immediately
1899 followed by that character. The same escapes apply to characters as
1900 to strings. So if you want to write the character backslash, you
1901 must write @kbd{'\\} where the first @code{\} escapes the second
1902 @code{\}. As you can see, the quote is an acute accent, not a
1903 grave accent. A newline
1905 @ifclear abnormal-separator
1906 (or semicolon @samp{;})
1908 @ifset abnormal-separator
1910 (or at sign @samp{@@})
1913 (or dollar sign @samp{$}, for the H8/300; or semicolon @samp{;} for the
1919 immediately following an acute accent is taken as a literal character
1920 and does not count as the end of a statement. The value of a character
1921 constant in a numeric expression is the machine's byte-wide code for
1922 that character. @code{@value{AS}} assumes your character code is ASCII:
1923 @kbd{'A} means 65, @kbd{'B} means 66, and so on. @refill
1926 @subsection Number Constants
1928 @cindex constants, number
1929 @cindex number constants
1930 @code{@value{AS}} distinguishes three kinds of numbers according to how they
1931 are stored in the target machine. @emph{Integers} are numbers that
1932 would fit into an @code{int} in the C language. @emph{Bignums} are
1933 integers, but they are stored in more than 32 bits. @emph{Flonums}
1934 are floating point numbers, described below.
1937 * Integers:: Integers
1942 * Bit Fields:: Bit Fields
1948 @subsubsection Integers
1950 @cindex constants, integer
1952 @cindex binary integers
1953 @cindex integers, binary
1954 A binary integer is @samp{0b} or @samp{0B} followed by zero or more of
1955 the binary digits @samp{01}.
1957 @cindex octal integers
1958 @cindex integers, octal
1959 An octal integer is @samp{0} followed by zero or more of the octal
1960 digits (@samp{01234567}).
1962 @cindex decimal integers
1963 @cindex integers, decimal
1964 A decimal integer starts with a non-zero digit followed by zero or
1965 more digits (@samp{0123456789}).
1967 @cindex hexadecimal integers
1968 @cindex integers, hexadecimal
1969 A hexadecimal integer is @samp{0x} or @samp{0X} followed by one or
1970 more hexadecimal digits chosen from @samp{0123456789abcdefABCDEF}.
1972 Integers have the usual values. To denote a negative integer, use
1973 the prefix operator @samp{-} discussed under expressions
1974 (@pxref{Prefix Ops,,Prefix Operators}).
1977 @subsubsection Bignums
1980 @cindex constants, bignum
1981 A @dfn{bignum} has the same syntax and semantics as an integer
1982 except that the number (or its negative) takes more than 32 bits to
1983 represent in binary. The distinction is made because in some places
1984 integers are permitted while bignums are not.
1987 @subsubsection Flonums
1989 @cindex floating point numbers
1990 @cindex constants, floating point
1992 @cindex precision, floating point
1993 A @dfn{flonum} represents a floating point number. The translation is
1994 indirect: a decimal floating point number from the text is converted by
1995 @code{@value{AS}} to a generic binary floating point number of more than
1996 sufficient precision. This generic floating point number is converted
1997 to a particular computer's floating point format (or formats) by a
1998 portion of @code{@value{AS}} specialized to that computer.
2000 A flonum is written by writing (in order)
2005 (@samp{0} is optional on the HPPA.)
2009 A letter, to tell @code{@value{AS}} the rest of the number is a flonum.
2011 @kbd{e} is recommended. Case is not important.
2013 @c FIXME: verify if flonum syntax really this vague for most cases
2014 (Any otherwise illegal letter works here, but that might be changed. Vax BSD
2015 4.2 assembler seems to allow any of @samp{defghDEFGH}.)
2018 On the H8/300, H8/500,
2020 and AMD 29K architectures, the letter must be
2021 one of the letters @samp{DFPRSX} (in upper or lower case).
2023 On the ARC, the letter must be one of the letters @samp{DFRS}
2024 (in upper or lower case).
2026 On the Intel 960 architecture, the letter must be
2027 one of the letters @samp{DFT} (in upper or lower case).
2029 On the HPPA architecture, the letter must be @samp{E} (upper case only).
2033 One of the letters @samp{DFPRSX} (in upper or lower case).
2036 One of the letters @samp{DFRS} (in upper or lower case).
2039 One of the letters @samp{DFPRSX} (in upper or lower case).
2042 The letter @samp{E} (upper case only).
2045 One of the letters @samp{DFT} (in upper or lower case).
2050 An optional sign: either @samp{+} or @samp{-}.
2053 An optional @dfn{integer part}: zero or more decimal digits.
2056 An optional @dfn{fractional part}: @samp{.} followed by zero
2057 or more decimal digits.
2060 An optional exponent, consisting of:
2064 An @samp{E} or @samp{e}.
2065 @c I can't find a config where "EXP_CHARS" is other than 'eE', but in
2066 @c principle this can perfectly well be different on different targets.
2068 Optional sign: either @samp{+} or @samp{-}.
2070 One or more decimal digits.
2075 At least one of the integer part or the fractional part must be
2076 present. The floating point number has the usual base-10 value.
2078 @code{@value{AS}} does all processing using integers. Flonums are computed
2079 independently of any floating point hardware in the computer running
2084 @c Bit fields are written as a general facility but are also controlled
2085 @c by a conditional-compilation flag---which is as of now (21mar91)
2086 @c turned on only by the i960 config of GAS.
2088 @subsubsection Bit Fields
2091 @cindex constants, bit field
2092 You can also define numeric constants as @dfn{bit fields}.
2093 specify two numbers separated by a colon---
2095 @var{mask}:@var{value}
2098 @code{@value{AS}} applies a bitwise @sc{and} between @var{mask} and
2101 The resulting number is then packed
2103 @c this conditional paren in case bit fields turned on elsewhere than 960
2104 (in host-dependent byte order)
2106 into a field whose width depends on which assembler directive has the
2107 bit-field as its argument. Overflow (a result from the bitwise and
2108 requiring more binary digits to represent) is not an error; instead,
2109 more constants are generated, of the specified width, beginning with the
2110 least significant digits.@refill
2112 The directives @code{.byte}, @code{.hword}, @code{.int}, @code{.long},
2113 @code{.short}, and @code{.word} accept bit-field arguments.
2118 @chapter Sections and Relocation
2123 * Secs Background:: Background
2124 * Ld Sections:: Linker Sections
2125 * As Sections:: Assembler Internal Sections
2126 * Sub-Sections:: Sub-Sections
2130 @node Secs Background
2133 Roughly, a section is a range of addresses, with no gaps; all data
2134 ``in'' those addresses is treated the same for some particular purpose.
2135 For example there may be a ``read only'' section.
2137 @cindex linker, and assembler
2138 @cindex assembler, and linker
2139 The linker @code{@value{LD}} reads many object files (partial programs) and
2140 combines their contents to form a runnable program. When @code{@value{AS}}
2141 emits an object file, the partial program is assumed to start at address 0.
2142 @code{@value{LD}} assigns the final addresses for the partial program, so that
2143 different partial programs do not overlap. This is actually an
2144 oversimplification, but it suffices to explain how @code{@value{AS}} uses
2147 @code{@value{LD}} moves blocks of bytes of your program to their run-time
2148 addresses. These blocks slide to their run-time addresses as rigid
2149 units; their length does not change and neither does the order of bytes
2150 within them. Such a rigid unit is called a @emph{section}. Assigning
2151 run-time addresses to sections is called @dfn{relocation}. It includes
2152 the task of adjusting mentions of object-file addresses so they refer to
2153 the proper run-time addresses.
2155 For the H8/300 and H8/500,
2156 and for the Hitachi SH,
2157 @code{@value{AS}} pads sections if needed to
2158 ensure they end on a word (sixteen bit) boundary.
2161 @cindex standard assembler sections
2162 An object file written by @code{@value{AS}} has at least three sections, any
2163 of which may be empty. These are named @dfn{text}, @dfn{data} and
2168 When it generates COFF output,
2170 @code{@value{AS}} can also generate whatever other named sections you specify
2171 using the @samp{.section} directive (@pxref{Section,,@code{.section}}).
2172 If you do not use any directives that place output in the @samp{.text}
2173 or @samp{.data} sections, these sections still exist, but are empty.
2178 When @code{@value{AS}} generates SOM or ELF output for the HPPA,
2180 @code{@value{AS}} can also generate whatever other named sections you
2181 specify using the @samp{.space} and @samp{.subspace} directives. See
2182 @cite{HP9000 Series 800 Assembly Language Reference Manual}
2183 (HP 92432-90001) for details on the @samp{.space} and @samp{.subspace}
2184 assembler directives.
2187 Additionally, @code{@value{AS}} uses different names for the standard
2188 text, data, and bss sections when generating SOM output. Program text
2189 is placed into the @samp{$CODE$} section, data into @samp{$DATA$}, and
2190 BSS into @samp{$BSS$}.
2194 Within the object file, the text section starts at address @code{0}, the
2195 data section follows, and the bss section follows the data section.
2198 When generating either SOM or ELF output files on the HPPA, the text
2199 section starts at address @code{0}, the data section at address
2200 @code{0x4000000}, and the bss section follows the data section.
2203 To let @code{@value{LD}} know which data changes when the sections are
2204 relocated, and how to change that data, @code{@value{AS}} also writes to the
2205 object file details of the relocation needed. To perform relocation
2206 @code{@value{LD}} must know, each time an address in the object
2210 Where in the object file is the beginning of this reference to
2213 How long (in bytes) is this reference?
2215 Which section does the address refer to? What is the numeric value of
2217 (@var{address}) @minus{} (@var{start-address of section})?
2220 Is the reference to an address ``Program-Counter relative''?
2223 @cindex addresses, format of
2224 @cindex section-relative addressing
2225 In fact, every address @code{@value{AS}} ever uses is expressed as
2227 (@var{section}) + (@var{offset into section})
2230 Further, most expressions @code{@value{AS}} computes have this section-relative
2233 (For some object formats, such as SOM for the HPPA, some expressions are
2234 symbol-relative instead.)
2237 In this manual we use the notation @{@var{secname} @var{N}@} to mean ``offset
2238 @var{N} into section @var{secname}.''
2240 Apart from text, data and bss sections you need to know about the
2241 @dfn{absolute} section. When @code{@value{LD}} mixes partial programs,
2242 addresses in the absolute section remain unchanged. For example, address
2243 @code{@{absolute 0@}} is ``relocated'' to run-time address 0 by
2244 @code{@value{LD}}. Although the linker never arranges two partial programs'
2245 data sections with overlapping addresses after linking, @emph{by definition}
2246 their absolute sections must overlap. Address @code{@{absolute@ 239@}} in one
2247 part of a program is always the same address when the program is running as
2248 address @code{@{absolute@ 239@}} in any other part of the program.
2250 The idea of sections is extended to the @dfn{undefined} section. Any
2251 address whose section is unknown at assembly time is by definition
2252 rendered @{undefined @var{U}@}---where @var{U} is filled in later.
2253 Since numbers are always defined, the only way to generate an undefined
2254 address is to mention an undefined symbol. A reference to a named
2255 common block would be such a symbol: its value is unknown at assembly
2256 time so it has section @emph{undefined}.
2258 By analogy the word @emph{section} is used to describe groups of sections in
2259 the linked program. @code{@value{LD}} puts all partial programs' text
2260 sections in contiguous addresses in the linked program. It is
2261 customary to refer to the @emph{text section} of a program, meaning all
2262 the addresses of all partial programs' text sections. Likewise for
2263 data and bss sections.
2265 Some sections are manipulated by @code{@value{LD}}; others are invented for
2266 use of @code{@value{AS}} and have no meaning except during assembly.
2269 @section Linker Sections
2270 @code{@value{LD}} deals with just four kinds of sections, summarized below.
2275 @cindex named sections
2276 @cindex sections, named
2277 @item named sections
2280 @cindex text section
2281 @cindex data section
2285 These sections hold your program. @code{@value{AS}} and @code{@value{LD}} treat them as
2286 separate but equal sections. Anything you can say of one section is
2289 When the program is running, however, it is
2290 customary for the text section to be unalterable. The
2291 text section is often shared among processes: it contains
2292 instructions, constants and the like. The data section of a running
2293 program is usually alterable: for example, C variables would be stored
2294 in the data section.
2299 This section contains zeroed bytes when your program begins running. It
2300 is used to hold unitialized variables or common storage. The length of
2301 each partial program's bss section is important, but because it starts
2302 out containing zeroed bytes there is no need to store explicit zero
2303 bytes in the object file. The bss section was invented to eliminate
2304 those explicit zeros from object files.
2306 @cindex absolute section
2307 @item absolute section
2308 Address 0 of this section is always ``relocated'' to runtime address 0.
2309 This is useful if you want to refer to an address that @code{@value{LD}} must
2310 not change when relocating. In this sense we speak of absolute
2311 addresses being ``unrelocatable'': they do not change during relocation.
2313 @cindex undefined section
2314 @item undefined section
2315 This ``section'' is a catch-all for address references to objects not in
2316 the preceding sections.
2317 @c FIXME: ref to some other doc on obj-file formats could go here.
2320 @cindex relocation example
2321 An idealized example of three relocatable sections follows.
2323 The example uses the traditional section names @samp{.text} and @samp{.data}.
2325 Memory addresses are on the horizontal axis.
2329 @c END TEXI2ROFF-KILL
2332 partial program # 1: |ttttt|dddd|00|
2339 partial program # 2: |TTT|DDD|000|
2342 +--+---+-----+--+----+---+-----+~~
2343 linked program: | |TTT|ttttt| |dddd|DDD|00000|
2344 +--+---+-----+--+----+---+-----+~~
2346 addresses: 0 @dots{}
2353 \line{\it Partial program \#1: \hfil}
2354 \line{\ibox{2.5cm}{\tt text}\ibox{2cm}{\tt data}\ibox{1cm}{\tt bss}\hfil}
2355 \line{\boxit{2.5cm}{\tt ttttt}\boxit{2cm}{\tt dddd}\boxit{1cm}{\tt 00}\hfil}
2357 \line{\it Partial program \#2: \hfil}
2358 \line{\ibox{1cm}{\tt text}\ibox{1.5cm}{\tt data}\ibox{1cm}{\tt bss}\hfil}
2359 \line{\boxit{1cm}{\tt TTT}\boxit{1.5cm}{\tt DDDD}\boxit{1cm}{\tt 000}\hfil}
2361 \line{\it linked program: \hfil}
2362 \line{\ibox{.5cm}{}\ibox{1cm}{\tt text}\ibox{2.5cm}{}\ibox{.75cm}{}\ibox{2cm}{\tt data}\ibox{1.5cm}{}\ibox{2cm}{\tt bss}\hfil}
2363 \line{\boxit{.5cm}{}\boxit{1cm}{\tt TTT}\boxit{2.5cm}{\tt
2364 ttttt}\boxit{.75cm}{}\boxit{2cm}{\tt dddd}\boxit{1.5cm}{\tt
2365 DDDD}\boxit{2cm}{\tt 00000}\ \dots\hfil}
2367 \line{\it addresses: \hfil}
2371 @c END TEXI2ROFF-KILL
2374 @section Assembler Internal Sections
2376 @cindex internal assembler sections
2377 @cindex sections in messages, internal
2378 These sections are meant only for the internal use of @code{@value{AS}}. They
2379 have no meaning at run-time. You do not really need to know about these
2380 sections for most purposes; but they can be mentioned in @code{@value{AS}}
2381 warning messages, so it might be helpful to have an idea of their
2382 meanings to @code{@value{AS}}. These sections are used to permit the
2383 value of every expression in your assembly language program to be a
2384 section-relative address.
2387 @cindex assembler internal logic error
2388 @item ASSEMBLER-INTERNAL-LOGIC-ERROR!
2389 An internal assembler logic error has been found. This means there is a
2390 bug in the assembler.
2392 @cindex expr (internal section)
2394 The assembler stores complex expression internally as combinations of
2395 symbols. When it needs to represent an expression as a symbol, it puts
2396 it in the expr section.
2398 @c FIXME item transfer[t] vector preload
2399 @c FIXME item transfer[t] vector postload
2400 @c FIXME item register
2404 @section Sub-Sections
2406 @cindex numbered subsections
2407 @cindex grouping data
2413 fall into two sections: text and data.
2415 You may have separate groups of
2417 data in named sections
2421 data in named sections
2427 that you want to end up near to each other in the object file, even though they
2428 are not contiguous in the assembler source. @code{@value{AS}} allows you to
2429 use @dfn{subsections} for this purpose. Within each section, there can be
2430 numbered subsections with values from 0 to 8192. Objects assembled into the
2431 same subsection go into the object file together with other objects in the same
2432 subsection. For example, a compiler might want to store constants in the text
2433 section, but might not want to have them interspersed with the program being
2434 assembled. In this case, the compiler could issue a @samp{.text 0} before each
2435 section of code being output, and a @samp{.text 1} before each group of
2436 constants being output.
2438 Subsections are optional. If you do not use subsections, everything
2439 goes in subsection number zero.
2442 Each subsection is zero-padded up to a multiple of four bytes.
2443 (Subsections may be padded a different amount on different flavors
2444 of @code{@value{AS}}.)
2448 On the H8/300 and H8/500 platforms, each subsection is zero-padded to a word
2449 boundary (two bytes).
2450 The same is true on the Hitachi SH.
2453 @c FIXME section padding (alignment)?
2454 @c Rich Pixley says padding here depends on target obj code format; that
2455 @c doesn't seem particularly useful to say without further elaboration,
2456 @c so for now I say nothing about it. If this is a generic BFD issue,
2457 @c these paragraphs might need to vanish from this manual, and be
2458 @c discussed in BFD chapter of binutils (or some such).
2461 On the AMD 29K family, no particular padding is added to section or
2462 subsection sizes; @value{AS} forces no alignment on this platform.
2466 Subsections appear in your object file in numeric order, lowest numbered
2467 to highest. (All this to be compatible with other people's assemblers.)
2468 The object file contains no representation of subsections; @code{@value{LD}} and
2469 other programs that manipulate object files see no trace of them.
2470 They just see all your text subsections as a text section, and all your
2471 data subsections as a data section.
2473 To specify which subsection you want subsequent statements assembled
2474 into, use a numeric argument to specify it, in a @samp{.text
2475 @var{expression}} or a @samp{.data @var{expression}} statement.
2478 When generating COFF output, you
2483 can also use an extra subsection
2484 argument with arbitrary named sections: @samp{.section @var{name},
2487 @var{Expression} should be an absolute expression.
2488 (@xref{Expressions}.) If you just say @samp{.text} then @samp{.text 0}
2489 is assumed. Likewise @samp{.data} means @samp{.data 0}. Assembly
2490 begins in @code{text 0}. For instance:
2492 .text 0 # The default subsection is text 0 anyway.
2493 .ascii "This lives in the first text subsection. *"
2495 .ascii "But this lives in the second text subsection."
2497 .ascii "This lives in the data section,"
2498 .ascii "in the first data subsection."
2500 .ascii "This lives in the first text section,"
2501 .ascii "immediately following the asterisk (*)."
2504 Each section has a @dfn{location counter} incremented by one for every byte
2505 assembled into that section. Because subsections are merely a convenience
2506 restricted to @code{@value{AS}} there is no concept of a subsection location
2507 counter. There is no way to directly manipulate a location counter---but the
2508 @code{.align} directive changes it, and any label definition captures its
2509 current value. The location counter of the section where statements are being
2510 assembled is said to be the @dfn{active} location counter.
2513 @section bss Section
2516 @cindex common variable storage
2517 The bss section is used for local common variable storage.
2518 You may allocate address space in the bss section, but you may
2519 not dictate data to load into it before your program executes. When
2520 your program starts running, all the contents of the bss
2521 section are zeroed bytes.
2523 The @code{.lcomm} pseudo-op defines a symbol in the bss section; see
2524 @ref{Lcomm,,@code{.lcomm}}.
2526 The @code{.comm} pseudo-op may be used to declare a common symbol, which is
2527 another form of uninitialized symbol; see @xref{Comm,,@code{.comm}}.
2530 When assembling for a target which supports multiple sections, such as ELF or
2531 COFF, you may switch into the @code{.bss} section and define symbols as usual;
2532 see @ref{Section,,@code{.section}}. You may only assemble zero values into the
2533 section. Typically the section will only contain symbol definitions and
2534 @code{.skip} directives (@pxref{Skip,,@code{.skip}}).
2541 Symbols are a central concept: the programmer uses symbols to name
2542 things, the linker uses symbols to link, and the debugger uses symbols
2546 @cindex debuggers, and symbol order
2547 @emph{Warning:} @code{@value{AS}} does not place symbols in the object file in
2548 the same order they were declared. This may break some debuggers.
2553 * Setting Symbols:: Giving Symbols Other Values
2554 * Symbol Names:: Symbol Names
2555 * Dot:: The Special Dot Symbol
2556 * Symbol Attributes:: Symbol Attributes
2563 A @dfn{label} is written as a symbol immediately followed by a colon
2564 @samp{:}. The symbol then represents the current value of the
2565 active location counter, and is, for example, a suitable instruction
2566 operand. You are warned if you use the same symbol to represent two
2567 different locations: the first definition overrides any other
2571 On the HPPA, the usual form for a label need not be immediately followed by a
2572 colon, but instead must start in column zero. Only one label may be defined on
2573 a single line. To work around this, the HPPA version of @code{@value{AS}} also
2574 provides a special directive @code{.label} for defining labels more flexibly.
2577 @node Setting Symbols
2578 @section Giving Symbols Other Values
2580 @cindex assigning values to symbols
2581 @cindex symbol values, assigning
2582 A symbol can be given an arbitrary value by writing a symbol, followed
2583 by an equals sign @samp{=}, followed by an expression
2584 (@pxref{Expressions}). This is equivalent to using the @code{.set}
2585 directive. @xref{Set,,@code{.set}}.
2588 @section Symbol Names
2590 @cindex symbol names
2591 @cindex names, symbol
2592 @ifclear SPECIAL-SYMS
2593 Symbol names begin with a letter or with one of @samp{._}. On most
2594 machines, you can also use @code{$} in symbol names; exceptions are
2595 noted in @ref{Machine Dependencies}. That character may be followed by any
2596 string of digits, letters, dollar signs (unless otherwise noted in
2597 @ref{Machine Dependencies}), and underscores.
2600 For the AMD 29K family, @samp{?} is also allowed in the
2601 body of a symbol name, though not at its beginning.
2606 Symbol names begin with a letter or with one of @samp{._}. On the
2608 H8/500, you can also use @code{$} in symbol names. That character may
2609 be followed by any string of digits, letters, dollar signs (save on the
2610 H8/300), and underscores.
2614 Case of letters is significant: @code{foo} is a different symbol name
2617 Each symbol has exactly one name. Each name in an assembly language program
2618 refers to exactly one symbol. You may use that symbol name any number of times
2621 @subheading Local Symbol Names
2623 @cindex local symbol names
2624 @cindex symbol names, local
2625 @cindex temporary symbol names
2626 @cindex symbol names, temporary
2627 Local symbols help compilers and programmers use names temporarily.
2628 There are ten local symbol names, which are re-used throughout the
2629 program. You may refer to them using the names @samp{0} @samp{1}
2630 @dots{} @samp{9}. To define a local symbol, write a label of the form
2631 @samp{@b{N}:} (where @b{N} represents any digit). To refer to the most
2632 recent previous definition of that symbol write @samp{@b{N}b}, using the
2633 same digit as when you defined the label. To refer to the next
2634 definition of a local label, write @samp{@b{N}f}---where @b{N} gives you
2635 a choice of 10 forward references. The @samp{b} stands for
2636 ``backwards'' and the @samp{f} stands for ``forwards''.
2638 Local symbols are not emitted by the current @sc{gnu} C compiler.
2640 There is no restriction on how you can use these labels, but
2641 remember that at any point in the assembly you can refer to at most
2642 10 prior local labels and to at most 10 forward local labels.
2644 Local symbol names are only a notation device. They are immediately
2645 transformed into more conventional symbol names before the assembler
2646 uses them. The symbol names stored in the symbol table, appearing in
2647 error messages and optionally emitted to the object file have these
2652 All local labels begin with @samp{L}. Normally both @code{@value{AS}} and
2653 @code{@value{LD}} forget symbols that start with @samp{L}. These labels are
2654 used for symbols you are never intended to see. If you use the
2655 @samp{-L} option then @code{@value{AS}} retains these symbols in the
2656 object file. If you also instruct @code{@value{LD}} to retain these symbols,
2657 you may use them in debugging.
2660 If the label is written @samp{0:} then the digit is @samp{0}.
2661 If the label is written @samp{1:} then the digit is @samp{1}.
2662 And so on up through @samp{9:}.
2665 This unusual character is included so you do not accidentally invent
2666 a symbol of the same name. The character has ASCII value
2669 @item @emph{ordinal number}
2670 This is a serial number to keep the labels distinct. The first
2671 @samp{0:} gets the number @samp{1}; The 15th @samp{0:} gets the
2672 number @samp{15}; @emph{etc.}. Likewise for the other labels @samp{1:}
2676 For instance, the first @code{1:} is named @code{L1@kbd{C-A}1}, the 44th
2677 @code{3:} is named @code{L3@kbd{C-A}44}.
2680 @section The Special Dot Symbol
2682 @cindex dot (symbol)
2683 @cindex @code{.} (symbol)
2684 @cindex current address
2685 @cindex location counter
2686 The special symbol @samp{.} refers to the current address that
2687 @code{@value{AS}} is assembling into. Thus, the expression @samp{melvin:
2688 .long .} defines @code{melvin} to contain its own address.
2689 Assigning a value to @code{.} is treated the same as a @code{.org}
2690 directive. Thus, the expression @samp{.=.+4} is the same as saying
2691 @ifclear no-space-dir
2700 @node Symbol Attributes
2701 @section Symbol Attributes
2703 @cindex symbol attributes
2704 @cindex attributes, symbol
2705 Every symbol has, as well as its name, the attributes ``Value'' and
2706 ``Type''. Depending on output format, symbols can also have auxiliary
2709 The detailed definitions are in @file{a.out.h}.
2712 If you use a symbol without defining it, @code{@value{AS}} assumes zero for
2713 all these attributes, and probably won't warn you. This makes the
2714 symbol an externally defined symbol, which is generally what you
2718 * Symbol Value:: Value
2719 * Symbol Type:: Type
2722 * a.out Symbols:: Symbol Attributes: @code{a.out}
2726 * a.out Symbols:: Symbol Attributes: @code{a.out}
2729 * a.out Symbols:: Symbol Attributes: @code{a.out}, @code{b.out}
2734 * COFF Symbols:: Symbol Attributes for COFF
2737 * SOM Symbols:: Symbol Attributes for SOM
2744 @cindex value of a symbol
2745 @cindex symbol value
2746 The value of a symbol is (usually) 32 bits. For a symbol which labels a
2747 location in the text, data, bss or absolute sections the value is the
2748 number of addresses from the start of that section to the label.
2749 Naturally for text, data and bss sections the value of a symbol changes
2750 as @code{@value{LD}} changes section base addresses during linking. Absolute
2751 symbols' values do not change during linking: that is why they are
2754 The value of an undefined symbol is treated in a special way. If it is
2755 0 then the symbol is not defined in this assembler source file, and
2756 @code{@value{LD}} tries to determine its value from other files linked into the
2757 same program. You make this kind of symbol simply by mentioning a symbol
2758 name without defining it. A non-zero value represents a @code{.comm}
2759 common declaration. The value is how much common storage to reserve, in
2760 bytes (addresses). The symbol refers to the first address of the
2766 @cindex type of a symbol
2768 The type attribute of a symbol contains relocation (section)
2769 information, any flag settings indicating that a symbol is external, and
2770 (optionally), other information for linkers and debuggers. The exact
2771 format depends on the object-code output format in use.
2776 @c The following avoids a "widow" subsection title. @group would be
2777 @c better if it were available outside examples.
2780 @subsection Symbol Attributes: @code{a.out}, @code{b.out}
2782 @cindex @code{b.out} symbol attributes
2783 @cindex symbol attributes, @code{b.out}
2784 These symbol attributes appear only when @code{@value{AS}} is configured for
2785 one of the Berkeley-descended object output formats---@code{a.out} or
2791 @subsection Symbol Attributes: @code{a.out}
2793 @cindex @code{a.out} symbol attributes
2794 @cindex symbol attributes, @code{a.out}
2800 @subsection Symbol Attributes: @code{a.out}
2802 @cindex @code{a.out} symbol attributes
2803 @cindex symbol attributes, @code{a.out}
2807 * Symbol Desc:: Descriptor
2808 * Symbol Other:: Other
2812 @subsubsection Descriptor
2814 @cindex descriptor, of @code{a.out} symbol
2815 This is an arbitrary 16-bit value. You may establish a symbol's
2816 descriptor value by using a @code{.desc} statement
2817 (@pxref{Desc,,@code{.desc}}). A descriptor value means nothing to
2821 @subsubsection Other
2823 @cindex other attribute, of @code{a.out} symbol
2824 This is an arbitrary 8-bit value. It means nothing to @code{@value{AS}}.
2829 @subsection Symbol Attributes for COFF
2831 @cindex COFF symbol attributes
2832 @cindex symbol attributes, COFF
2834 The COFF format supports a multitude of auxiliary symbol attributes;
2835 like the primary symbol attributes, they are set between @code{.def} and
2836 @code{.endef} directives.
2838 @subsubsection Primary Attributes
2840 @cindex primary attributes, COFF symbols
2841 The symbol name is set with @code{.def}; the value and type,
2842 respectively, with @code{.val} and @code{.type}.
2844 @subsubsection Auxiliary Attributes
2846 @cindex auxiliary attributes, COFF symbols
2847 The @code{@value{AS}} directives @code{.dim}, @code{.line}, @code{.scl},
2848 @code{.size}, and @code{.tag} can generate auxiliary symbol table
2849 information for COFF.
2854 @subsection Symbol Attributes for SOM
2856 @cindex SOM symbol attributes
2857 @cindex symbol attributes, SOM
2859 The SOM format for the HPPA supports a multitude of symbol attributes set with
2860 the @code{.EXPORT} and @code{.IMPORT} directives.
2862 The attributes are described in @cite{HP9000 Series 800 Assembly
2863 Language Reference Manual} (HP 92432-90001) under the @code{IMPORT} and
2864 @code{EXPORT} assembler directive documentation.
2868 @chapter Expressions
2872 @cindex numeric values
2873 An @dfn{expression} specifies an address or numeric value.
2874 Whitespace may precede and/or follow an expression.
2876 The result of an expression must be an absolute number, or else an offset into
2877 a particular section. If an expression is not absolute, and there is not
2878 enough information when @code{@value{AS}} sees the expression to know its
2879 section, a second pass over the source program might be necessary to interpret
2880 the expression---but the second pass is currently not implemented.
2881 @code{@value{AS}} aborts with an error message in this situation.
2884 * Empty Exprs:: Empty Expressions
2885 * Integer Exprs:: Integer Expressions
2889 @section Empty Expressions
2891 @cindex empty expressions
2892 @cindex expressions, empty
2893 An empty expression has no value: it is just whitespace or null.
2894 Wherever an absolute expression is required, you may omit the
2895 expression, and @code{@value{AS}} assumes a value of (absolute) 0. This
2896 is compatible with other assemblers.
2899 @section Integer Expressions
2901 @cindex integer expressions
2902 @cindex expressions, integer
2903 An @dfn{integer expression} is one or more @emph{arguments} delimited
2904 by @emph{operators}.
2907 * Arguments:: Arguments
2908 * Operators:: Operators
2909 * Prefix Ops:: Prefix Operators
2910 * Infix Ops:: Infix Operators
2914 @subsection Arguments
2916 @cindex expression arguments
2917 @cindex arguments in expressions
2918 @cindex operands in expressions
2919 @cindex arithmetic operands
2920 @dfn{Arguments} are symbols, numbers or subexpressions. In other
2921 contexts arguments are sometimes called ``arithmetic operands''. In
2922 this manual, to avoid confusing them with the ``instruction operands'' of
2923 the machine language, we use the term ``argument'' to refer to parts of
2924 expressions only, reserving the word ``operand'' to refer only to machine
2925 instruction operands.
2927 Symbols are evaluated to yield @{@var{section} @var{NNN}@} where
2928 @var{section} is one of text, data, bss, absolute,
2929 or undefined. @var{NNN} is a signed, 2's complement 32 bit
2932 Numbers are usually integers.
2934 A number can be a flonum or bignum. In this case, you are warned
2935 that only the low order 32 bits are used, and @code{@value{AS}} pretends
2936 these 32 bits are an integer. You may write integer-manipulating
2937 instructions that act on exotic constants, compatible with other
2940 @cindex subexpressions
2941 Subexpressions are a left parenthesis @samp{(} followed by an integer
2942 expression, followed by a right parenthesis @samp{)}; or a prefix
2943 operator followed by an argument.
2946 @subsection Operators
2948 @cindex operators, in expressions
2949 @cindex arithmetic functions
2950 @cindex functions, in expressions
2951 @dfn{Operators} are arithmetic functions, like @code{+} or @code{%}. Prefix
2952 operators are followed by an argument. Infix operators appear
2953 between their arguments. Operators may be preceded and/or followed by
2957 @subsection Prefix Operator
2959 @cindex prefix operators
2960 @code{@value{AS}} has the following @dfn{prefix operators}. They each take
2961 one argument, which must be absolute.
2963 @c the tex/end tex stuff surrounding this small table is meant to make
2964 @c it align, on the printed page, with the similar table in the next
2965 @c section (which is inside an enumerate).
2967 \global\advance\leftskip by \itemindent
2972 @dfn{Negation}. Two's complement negation.
2974 @dfn{Complementation}. Bitwise not.
2978 \global\advance\leftskip by -\itemindent
2982 @subsection Infix Operators
2984 @cindex infix operators
2985 @cindex operators, permitted arguments
2986 @dfn{Infix operators} take two arguments, one on either side. Operators
2987 have precedence, but operations with equal precedence are performed left
2988 to right. Apart from @code{+} or @code{-}, both arguments must be
2989 absolute, and the result is absolute.
2992 @cindex operator precedence
2993 @cindex precedence of operators
3000 @dfn{Multiplication}.
3003 @dfn{Division}. Truncation is the same as the C operator @samp{/}
3010 @dfn{Shift Left}. Same as the C operator @samp{<<}.
3014 @dfn{Shift Right}. Same as the C operator @samp{>>}.
3018 Intermediate precedence
3023 @dfn{Bitwise Inclusive Or}.
3029 @dfn{Bitwise Exclusive Or}.
3032 @dfn{Bitwise Or Not}.
3039 @cindex addition, permitted arguments
3040 @cindex plus, permitted arguments
3041 @cindex arguments for addition
3043 @dfn{Addition}. If either argument is absolute, the result has the section of
3044 the other argument. You may not add together arguments from different
3047 @cindex subtraction, permitted arguments
3048 @cindex minus, permitted arguments
3049 @cindex arguments for subtraction
3051 @dfn{Subtraction}. If the right argument is absolute, the
3052 result has the section of the left argument.
3053 If both arguments are in the same section, the result is absolute.
3054 You may not subtract arguments from different sections.
3055 @c FIXME is there still something useful to say about undefined - undefined ?
3059 In short, it's only meaningful to add or subtract the @emph{offsets} in an
3060 address; you can only have a defined section in one of the two arguments.
3063 @chapter Assembler Directives
3065 @cindex directives, machine independent
3066 @cindex pseudo-ops, machine independent
3067 @cindex machine independent directives
3068 All assembler directives have names that begin with a period (@samp{.}).
3069 The rest of the name is letters, usually in lower case.
3071 This chapter discusses directives that are available regardless of the
3072 target machine configuration for the @sc{gnu} assembler.
3074 Some machine configurations provide additional directives.
3075 @xref{Machine Dependencies}.
3078 @ifset machine-directives
3079 @xref{Machine Dependencies} for additional directives.
3084 * Abort:: @code{.abort}
3086 * ABORT:: @code{.ABORT}
3089 * Align:: @code{.align @var{abs-expr} , @var{abs-expr}}
3090 * Ascii:: @code{.ascii "@var{string}"}@dots{}
3091 * Asciz:: @code{.asciz "@var{string}"}@dots{}
3092 * Balign:: @code{.balign @var{abs-expr} , @var{abs-expr}}
3093 * Byte:: @code{.byte @var{expressions}}
3094 * Comm:: @code{.comm @var{symbol} , @var{length} }
3095 * Data:: @code{.data @var{subsection}}
3097 * Def:: @code{.def @var{name}}
3100 * Desc:: @code{.desc @var{symbol}, @var{abs-expression}}
3106 * Double:: @code{.double @var{flonums}}
3107 * Eject:: @code{.eject}
3108 * Else:: @code{.else}
3111 * Endef:: @code{.endef}
3114 * Endfunc:: @code{.endfunc}
3115 * Endif:: @code{.endif}
3116 * Equ:: @code{.equ @var{symbol}, @var{expression}}
3117 * Equiv:: @code{.equiv @var{symbol}, @var{expression}}
3119 * Exitm:: @code{.exitm}
3120 * Extern:: @code{.extern}
3121 * Fail:: @code{.fail}
3122 @ifclear no-file-dir
3123 * File:: @code{.file @var{string}}
3126 * Fill:: @code{.fill @var{repeat} , @var{size} , @var{value}}
3127 * Float:: @code{.float @var{flonums}}
3128 * Func:: @code{.func}
3129 * Global:: @code{.global @var{symbol}}, @code{.globl @var{symbol}}
3130 * hword:: @code{.hword @var{expressions}}
3131 * Ident:: @code{.ident}
3132 * If:: @code{.if @var{absolute expression}}
3133 * Include:: @code{.include "@var{file}"}
3134 * Int:: @code{.int @var{expressions}}
3135 * Irp:: @code{.irp @var{symbol},@var{values}}@dots{}
3136 * Irpc:: @code{.irpc @var{symbol},@var{values}}@dots{}
3137 * Lcomm:: @code{.lcomm @var{symbol} , @var{length}}
3138 * Lflags:: @code{.lflags}
3139 @ifclear no-line-dir
3140 * Line:: @code{.line @var{line-number}}
3143 * Ln:: @code{.ln @var{line-number}}
3144 * Linkonce:: @code{.linkonce [@var{type}]}
3145 * List:: @code{.list}
3146 * Long:: @code{.long @var{expressions}}
3148 * Lsym:: @code{.lsym @var{symbol}, @var{expression}}
3151 * Macro:: @code{.macro @var{name} @var{args}}@dots{}
3152 * MRI:: @code{.mri @var{val}}
3154 * Nolist:: @code{.nolist}
3155 * Octa:: @code{.octa @var{bignums}}
3156 * Org:: @code{.org @var{new-lc} , @var{fill}}
3157 * P2align:: @code{.p2align @var{abs-expr} , @var{abs-expr}}
3158 * Print:: @code{.print @var{string}}
3159 * Psize:: @code{.psize @var{lines}, @var{columns}}
3160 * Purgem:: @code{.purgem @var{name}}
3161 * Quad:: @code{.quad @var{bignums}}
3162 * Rept:: @code{.rept @var{count}}
3163 * Sbttl:: @code{.sbttl "@var{subheading}"}
3165 * Scl:: @code{.scl @var{class}}
3166 * Section:: @code{.section @var{name}, @var{subsection}}
3169 * Set:: @code{.set @var{symbol}, @var{expression}}
3170 * Short:: @code{.short @var{expressions}}
3171 * Single:: @code{.single @var{flonums}}
3173 * Size:: @code{.size}
3176 * Skip:: @code{.skip @var{size} , @var{fill}}
3177 * Sleb128:: @code{.sleb128 @var{expressions}}
3178 * Space:: @code{.space @var{size} , @var{fill}}
3180 * Stab:: @code{.stabd, .stabn, .stabs}
3183 * String:: @code{.string "@var{str}"}
3184 * Struct:: @code{.struct @var{expression}}
3186 * Symver:: @code{.symver @var{name},@var{name2@@nodename}}
3189 * Tag:: @code{.tag @var{structname}}
3192 * Text:: @code{.text @var{subsection}}
3193 * Title:: @code{.title "@var{heading}"}
3195 * Type:: @code{.type @var{int}}
3196 * Val:: @code{.val @var{addr}}
3199 * Uleb128:: @code{.uleb128 @var{expressions}}
3200 * Word:: @code{.word @var{expressions}}
3201 * Deprecated:: Deprecated Directives
3205 @section @code{.abort}
3207 @cindex @code{abort} directive
3208 @cindex stopping the assembly
3209 This directive stops the assembly immediately. It is for
3210 compatibility with other assemblers. The original idea was that the
3211 assembly language source would be piped into the assembler. If the sender
3212 of the source quit, it could use this directive tells @code{@value{AS}} to
3213 quit also. One day @code{.abort} will not be supported.
3217 @section @code{.ABORT}
3219 @cindex @code{ABORT} directive
3220 When producing COFF output, @code{@value{AS}} accepts this directive as a
3221 synonym for @samp{.abort}.
3224 When producing @code{b.out} output, @code{@value{AS}} accepts this directive,
3230 @section @code{.align @var{abs-expr}, @var{abs-expr}, @var{abs-expr}}
3232 @cindex padding the location counter
3233 @cindex @code{align} directive
3234 Pad the location counter (in the current subsection) to a particular storage
3235 boundary. The first expression (which must be absolute) is the alignment
3236 required, as described below.
3238 The second expression (also absolute) gives the fill value to be stored in the
3239 padding bytes. It (and the comma) may be omitted. If it is omitted, the
3240 padding bytes are normally zero. However, on some systems, if the section is
3241 marked as containing code and the fill value is omitted, the space is filled
3242 with no-op instructions.
3244 The third expression is also absolute, and is also optional. If it is present,
3245 it is the maximum number of bytes that should be skipped by this alignment
3246 directive. If doing the alignment would require skipping more bytes than the
3247 specified maximum, then the alignment is not done at all. You can omit the
3248 fill value (the second argument) entirely by simply using two commas after the
3249 required alignment; this can be useful if you want the alignment to be filled
3250 with no-op instructions when appropriate.
3252 The way the required alignment is specified varies from system to system.
3253 For the a29k, hppa, m68k, m88k, w65, sparc, and Hitachi SH, and i386 using ELF
3255 the first expression is the
3256 alignment request in bytes. For example @samp{.align 8} advances
3257 the location counter until it is a multiple of 8. If the location counter
3258 is already a multiple of 8, no change is needed.
3260 For other systems, including the i386 using a.out format, it is the
3261 number of low-order zero bits the location counter must have after
3262 advancement. For example @samp{.align 3} advances the location
3263 counter until it a multiple of 8. If the location counter is already a
3264 multiple of 8, no change is needed.
3266 This inconsistency is due to the different behaviors of the various
3267 native assemblers for these systems which GAS must emulate.
3268 GAS also provides @code{.balign} and @code{.p2align} directives,
3269 described later, which have a consistent behavior across all
3270 architectures (but are specific to GAS).
3273 @section @code{.ascii "@var{string}"}@dots{}
3275 @cindex @code{ascii} directive
3276 @cindex string literals
3277 @code{.ascii} expects zero or more string literals (@pxref{Strings})
3278 separated by commas. It assembles each string (with no automatic
3279 trailing zero byte) into consecutive addresses.
3282 @section @code{.asciz "@var{string}"}@dots{}
3284 @cindex @code{asciz} directive
3285 @cindex zero-terminated strings
3286 @cindex null-terminated strings
3287 @code{.asciz} is just like @code{.ascii}, but each string is followed by
3288 a zero byte. The ``z'' in @samp{.asciz} stands for ``zero''.
3291 @section @code{.balign[wl] @var{abs-expr}, @var{abs-expr}, @var{abs-expr}}
3293 @cindex padding the location counter given number of bytes
3294 @cindex @code{balign} directive
3295 Pad the location counter (in the current subsection) to a particular
3296 storage boundary. The first expression (which must be absolute) is the
3297 alignment request in bytes. For example @samp{.balign 8} advances
3298 the location counter until it is a multiple of 8. If the location counter
3299 is already a multiple of 8, no change is needed.
3301 The second expression (also absolute) gives the fill value to be stored in the
3302 padding bytes. It (and the comma) may be omitted. If it is omitted, the
3303 padding bytes are normally zero. However, on some systems, if the section is
3304 marked as containing code and the fill value is omitted, the space is filled
3305 with no-op instructions.
3307 The third expression is also absolute, and is also optional. If it is present,
3308 it is the maximum number of bytes that should be skipped by this alignment
3309 directive. If doing the alignment would require skipping more bytes than the
3310 specified maximum, then the alignment is not done at all. You can omit the
3311 fill value (the second argument) entirely by simply using two commas after the
3312 required alignment; this can be useful if you want the alignment to be filled
3313 with no-op instructions when appropriate.
3315 @cindex @code{balignw} directive
3316 @cindex @code{balignl} directive
3317 The @code{.balignw} and @code{.balignl} directives are variants of the
3318 @code{.balign} directive. The @code{.balignw} directive treats the fill
3319 pattern as a two byte word value. The @code{.balignl} directives treats the
3320 fill pattern as a four byte longword value. For example, @code{.balignw
3321 4,0x368d} will align to a multiple of 4. If it skips two bytes, they will be
3322 filled in with the value 0x368d (the exact placement of the bytes depends upon
3323 the endianness of the processor). If it skips 1 or 3 bytes, the fill value is
3327 @section @code{.byte @var{expressions}}
3329 @cindex @code{byte} directive
3330 @cindex integers, one byte
3331 @code{.byte} expects zero or more expressions, separated by commas.
3332 Each expression is assembled into the next byte.
3335 @section @code{.comm @var{symbol} , @var{length} }
3337 @cindex @code{comm} directive
3338 @cindex symbol, common
3339 @code{.comm} declares a common symbol named @var{symbol}. When linking, a
3340 common symbol in one object file may be merged with a defined or common symbol
3341 of the same name in another object file. If @code{@value{LD}} does not see a
3342 definition for the symbol--just one or more common symbols--then it will
3343 allocate @var{length} bytes of uninitialized memory. @var{length} must be an
3344 absolute expression. If @code{@value{LD}} sees multiple common symbols with
3345 the same name, and they do not all have the same size, it will allocate space
3346 using the largest size.
3349 When using ELF, the @code{.comm} directive takes an optional third argument.
3350 This is the desired alignment of the symbol, specified as a byte boundary (for
3351 example, an alignment of 16 means that the least significant 4 bits of the
3352 address should be zero). The alignment must be an absolute expression, and it
3353 must be a power of two. If @code{@value{LD}} allocates uninitialized memory
3354 for the common symbol, it will use the alignment when placing the symbol. If
3355 no alignment is specified, @code{@value{AS}} will set the alignment to the
3356 largest power of two less than or equal to the size of the symbol, up to a
3361 The syntax for @code{.comm} differs slightly on the HPPA. The syntax is
3362 @samp{@var{symbol} .comm, @var{length}}; @var{symbol} is optional.
3366 @section @code{.data @var{subsection}}
3368 @cindex @code{data} directive
3369 @code{.data} tells @code{@value{AS}} to assemble the following statements onto the
3370 end of the data subsection numbered @var{subsection} (which is an
3371 absolute expression). If @var{subsection} is omitted, it defaults
3376 @section @code{.def @var{name}}
3378 @cindex @code{def} directive
3379 @cindex COFF symbols, debugging
3380 @cindex debugging COFF symbols
3381 Begin defining debugging information for a symbol @var{name}; the
3382 definition extends until the @code{.endef} directive is encountered.
3385 This directive is only observed when @code{@value{AS}} is configured for COFF
3386 format output; when producing @code{b.out}, @samp{.def} is recognized,
3393 @section @code{.desc @var{symbol}, @var{abs-expression}}
3395 @cindex @code{desc} directive
3396 @cindex COFF symbol descriptor
3397 @cindex symbol descriptor, COFF
3398 This directive sets the descriptor of the symbol (@pxref{Symbol Attributes})
3399 to the low 16 bits of an absolute expression.
3402 The @samp{.desc} directive is not available when @code{@value{AS}} is
3403 configured for COFF output; it is only for @code{a.out} or @code{b.out}
3404 object format. For the sake of compatibility, @code{@value{AS}} accepts
3405 it, but produces no output, when configured for COFF.
3411 @section @code{.dim}
3413 @cindex @code{dim} directive
3414 @cindex COFF auxiliary symbol information
3415 @cindex auxiliary symbol information, COFF
3416 This directive is generated by compilers to include auxiliary debugging
3417 information in the symbol table. It is only permitted inside
3418 @code{.def}/@code{.endef} pairs.
3421 @samp{.dim} is only meaningful when generating COFF format output; when
3422 @code{@value{AS}} is generating @code{b.out}, it accepts this directive but
3428 @section @code{.double @var{flonums}}
3430 @cindex @code{double} directive
3431 @cindex floating point numbers (double)
3432 @code{.double} expects zero or more flonums, separated by commas. It
3433 assembles floating point numbers.
3435 The exact kind of floating point numbers emitted depends on how
3436 @code{@value{AS}} is configured. @xref{Machine Dependencies}.
3440 On the @value{TARGET} family @samp{.double} emits 64-bit floating-point numbers
3441 in @sc{ieee} format.
3446 @section @code{.eject}
3448 @cindex @code{eject} directive
3449 @cindex new page, in listings
3450 @cindex page, in listings
3451 @cindex listing control: new page
3452 Force a page break at this point, when generating assembly listings.
3455 @section @code{.else}
3457 @cindex @code{else} directive
3458 @code{.else} is part of the @code{@value{AS}} support for conditional
3459 assembly; @pxref{If,,@code{.if}}. It marks the beginning of a section
3460 of code to be assembled if the condition for the preceding @code{.if}
3464 @section @code{.end}
3466 @cindex @code{end} directive
3467 @code{.end} marks the end of the assembly file. @code{@value{AS}} does not
3468 process anything in the file past the @code{.end} directive.
3472 @section @code{.endef}
3474 @cindex @code{endef} directive
3475 This directive flags the end of a symbol definition begun with
3479 @samp{.endef} is only meaningful when generating COFF format output; if
3480 @code{@value{AS}} is configured to generate @code{b.out}, it accepts this
3481 directive but ignores it.
3486 @section @code{.endfunc}
3487 @cindex @code{endfunc} directive
3488 @code{.endfunc} marks the end of a function specified with @code{.func}.
3491 @section @code{.endif}
3493 @cindex @code{endif} directive
3494 @code{.endif} is part of the @code{@value{AS}} support for conditional assembly;
3495 it marks the end of a block of code that is only assembled
3496 conditionally. @xref{If,,@code{.if}}.
3499 @section @code{.equ @var{symbol}, @var{expression}}
3501 @cindex @code{equ} directive
3502 @cindex assigning values to symbols
3503 @cindex symbols, assigning values to
3504 This directive sets the value of @var{symbol} to @var{expression}.
3505 It is synonymous with @samp{.set}; @pxref{Set,,@code{.set}}.
3508 The syntax for @code{equ} on the HPPA is
3509 @samp{@var{symbol} .equ @var{expression}}.
3513 @section @code{.equiv @var{symbol}, @var{expression}}
3514 @cindex @code{equiv} directive
3515 The @code{.equiv} directive is like @code{.equ} and @code{.set}, except that
3516 the assembler will signal an error if @var{symbol} is already defined.
3518 Except for the contents of the error message, this is roughly equivalent to
3527 @section @code{.err}
3528 @cindex @code{err} directive
3529 If @code{@value{AS}} assembles a @code{.err} directive, it will print an error
3530 message and, unless the @code{-Z} option was used, it will not generate an
3531 object file. This can be used to signal error an conditionally compiled code.
3534 @section @code{.exitm}
3535 Exit early from the current macro definition. @xref{Macro}.
3538 @section @code{.extern}
3540 @cindex @code{extern} directive
3541 @code{.extern} is accepted in the source program---for compatibility
3542 with other assemblers---but it is ignored. @code{@value{AS}} treats
3543 all undefined symbols as external.
3546 @section @code{.fail @var{expression}}
3548 @cindex @code{fail} directive
3549 Generates an error or a warning. If the value of the @var{expression} is 500
3550 or more, @code{@value{AS}} will print a warning message. If the value is less
3551 than 500, @code{@value{AS}} will print an error message. The message will
3552 include the value of @var{expression}. This can occasionally be useful inside
3553 complex nested macros or conditional assembly.
3555 @ifclear no-file-dir
3557 @section @code{.file @var{string}}
3559 @cindex @code{file} directive
3560 @cindex logical file name
3561 @cindex file name, logical
3562 @code{.file} tells @code{@value{AS}} that we are about to start a new logical
3563 file. @var{string} is the new file name. In general, the filename is
3564 recognized whether or not it is surrounded by quotes @samp{"}; but if you wish
3565 to specify an empty file name, you must give the quotes--@code{""}. This
3566 statement may go away in future: it is only recognized to be compatible with
3567 old @code{@value{AS}} programs.
3569 In some configurations of @code{@value{AS}}, @code{.file} has already been
3570 removed to avoid conflicts with other assemblers. @xref{Machine Dependencies}.
3575 @section @code{.fill @var{repeat} , @var{size} , @var{value}}
3577 @cindex @code{fill} directive
3578 @cindex writing patterns in memory
3579 @cindex patterns, writing in memory
3580 @var{result}, @var{size} and @var{value} are absolute expressions.
3581 This emits @var{repeat} copies of @var{size} bytes. @var{Repeat}
3582 may be zero or more. @var{Size} may be zero or more, but if it is
3583 more than 8, then it is deemed to have the value 8, compatible with
3584 other people's assemblers. The contents of each @var{repeat} bytes
3585 is taken from an 8-byte number. The highest order 4 bytes are
3586 zero. The lowest order 4 bytes are @var{value} rendered in the
3587 byte-order of an integer on the computer @code{@value{AS}} is assembling for.
3588 Each @var{size} bytes in a repetition is taken from the lowest order
3589 @var{size} bytes of this number. Again, this bizarre behavior is
3590 compatible with other people's assemblers.
3592 @var{size} and @var{value} are optional.
3593 If the second comma and @var{value} are absent, @var{value} is
3594 assumed zero. If the first comma and following tokens are absent,
3595 @var{size} is assumed to be 1.
3598 @section @code{.float @var{flonums}}
3600 @cindex floating point numbers (single)
3601 @cindex @code{float} directive
3602 This directive assembles zero or more flonums, separated by commas. It
3603 has the same effect as @code{.single}.
3605 The exact kind of floating point numbers emitted depends on how
3606 @code{@value{AS}} is configured.
3607 @xref{Machine Dependencies}.
3611 On the @value{TARGET} family, @code{.float} emits 32-bit floating point numbers
3612 in @sc{ieee} format.
3617 @section @code{.func @var{name}[,@var{label}]}
3618 @cindex @code{func} directive
3619 @code{.func} emits debugging information to denote function @var{name}, and
3620 is ignored unless the file is assembled with debugging enabled.
3621 Only @samp{--gstabs} is currently supported.
3622 @var{label} is the entry point of the function and if omitted @var{name}
3623 prepended with the @samp{leading char} is used.
3624 @samp{leading char} is usually @code{_} or nothing, depending on the target.
3625 All functions are currently defined to have @code{void} return type.
3626 The function must be terminated with @code{.endfunc}.
3629 @section @code{.global @var{symbol}}, @code{.globl @var{symbol}}
3631 @cindex @code{global} directive
3632 @cindex symbol, making visible to linker
3633 @code{.global} makes the symbol visible to @code{@value{LD}}. If you define
3634 @var{symbol} in your partial program, its value is made available to
3635 other partial programs that are linked with it. Otherwise,
3636 @var{symbol} takes its attributes from a symbol of the same name
3637 from another file linked into the same program.
3639 Both spellings (@samp{.globl} and @samp{.global}) are accepted, for
3640 compatibility with other assemblers.
3643 On the HPPA, @code{.global} is not always enough to make it accessible to other
3644 partial programs. You may need the HPPA-only @code{.EXPORT} directive as well.
3645 @xref{HPPA Directives,, HPPA Assembler Directives}.
3649 @section @code{.hword @var{expressions}}
3651 @cindex @code{hword} directive
3652 @cindex integers, 16-bit
3653 @cindex numbers, 16-bit
3654 @cindex sixteen bit integers
3655 This expects zero or more @var{expressions}, and emits
3656 a 16 bit number for each.
3659 This directive is a synonym for @samp{.short}; depending on the target
3660 architecture, it may also be a synonym for @samp{.word}.
3664 This directive is a synonym for @samp{.short}.
3667 This directive is a synonym for both @samp{.short} and @samp{.word}.
3672 @section @code{.ident}
3674 @cindex @code{ident} directive
3675 This directive is used by some assemblers to place tags in object files.
3676 @code{@value{AS}} simply accepts the directive for source-file
3677 compatibility with such assemblers, but does not actually emit anything
3681 @section @code{.if @var{absolute expression}}
3683 @cindex conditional assembly
3684 @cindex @code{if} directive
3685 @code{.if} marks the beginning of a section of code which is only
3686 considered part of the source program being assembled if the argument
3687 (which must be an @var{absolute expression}) is non-zero. The end of
3688 the conditional section of code must be marked by @code{.endif}
3689 (@pxref{Endif,,@code{.endif}}); optionally, you may include code for the
3690 alternative condition, flagged by @code{.else} (@pxref{Else,,@code{.else}}).
3692 The following variants of @code{.if} are also supported:
3694 @cindex @code{ifdef} directive
3695 @item .ifdef @var{symbol}
3696 Assembles the following section of code if the specified @var{symbol}
3699 @cindex @code{ifc} directive
3700 @item .ifc @var{string1},@var{string2}
3701 Assembles the following section of code if the two strings are the same. The
3702 strings may be optionally quoted with single quotes. If they are not quoted,
3703 the first string stops at the first comma, and the second string stops at the
3704 end of the line. Strings which contain whitespace should be quoted. The
3705 string comparison is case sensitive.
3707 @cindex @code{ifeq} directive
3708 @item .ifeq @var{absolute expression}
3709 Assembles the following section of code if the argument is zero.
3711 @cindex @code{ifeqs} directive
3712 @item .ifeqs @var{string1},@var{string2}
3713 Another form of @code{.ifc}. The strings must be quoted using double quotes.
3715 @cindex @code{ifge} directive
3716 @item .ifge @var{absolute expression}
3717 Assembles the following section of code if the argument is greater than or
3720 @cindex @code{ifgt} directive
3721 @item .ifgt @var{absolute expression}
3722 Assembles the following section of code if the argument is greater than zero.
3724 @cindex @code{ifle} directive
3725 @item .ifle @var{absolute expression}
3726 Assembles the following section of code if the argument is less than or equal
3729 @cindex @code{iflt} directive
3730 @item .iflt @var{absolute expression}
3731 Assembles the following section of code if the argument is less than zero.
3733 @cindex @code{ifnc} directive
3734 @item .ifnc @var{string1},@var{string2}.
3735 Like @code{.ifc}, but the sense of the test is reversed: this assembles the
3736 following section of code if the two strings are not the same.
3738 @cindex @code{ifndef} directive
3739 @cindex @code{ifnotdef} directive
3740 @item .ifndef @var{symbol}
3741 @itemx .ifnotdef @var{symbol}
3742 Assembles the following section of code if the specified @var{symbol}
3743 has not been defined. Both spelling variants are equivalent.
3745 @cindex @code{ifne} directive
3746 @item .ifne @var{absolute expression}
3747 Assembles the following section of code if the argument is not equal to zero
3748 (in other words, this is equivalent to @code{.if}).
3750 @cindex @code{ifnes} directive
3751 @item .ifnes @var{string1},@var{string2}
3752 Like @code{.ifeqs}, but the sense of the test is reversed: this assembles the
3753 following section of code if the two strings are not the same.
3757 @section @code{.include "@var{file}"}
3759 @cindex @code{include} directive
3760 @cindex supporting files, including
3761 @cindex files, including
3762 This directive provides a way to include supporting files at specified
3763 points in your source program. The code from @var{file} is assembled as
3764 if it followed the point of the @code{.include}; when the end of the
3765 included file is reached, assembly of the original file continues. You
3766 can control the search paths used with the @samp{-I} command-line option
3767 (@pxref{Invoking,,Command-Line Options}). Quotation marks are required
3771 @section @code{.int @var{expressions}}
3773 @cindex @code{int} directive
3774 @cindex integers, 32-bit
3775 Expect zero or more @var{expressions}, of any section, separated by commas.
3776 For each expression, emit a number that, at run time, is the value of that
3777 expression. The byte order and bit size of the number depends on what kind
3778 of target the assembly is for.
3782 On the H8/500 and most forms of the H8/300, @code{.int} emits 16-bit
3783 integers. On the H8/300H and the Hitachi SH, however, @code{.int} emits
3789 @section @code{.irp @var{symbol},@var{values}}@dots{}
3791 @cindex @code{irp} directive
3792 Evaluate a sequence of statements assigning different values to @var{symbol}.
3793 The sequence of statements starts at the @code{.irp} directive, and is
3794 terminated by an @code{.endr} directive. For each @var{value}, @var{symbol} is
3795 set to @var{value}, and the sequence of statements is assembled. If no
3796 @var{value} is listed, the sequence of statements is assembled once, with
3797 @var{symbol} set to the null string. To refer to @var{symbol} within the
3798 sequence of statements, use @var{\symbol}.
3800 For example, assembling
3808 is equivalent to assembling
3817 @section @code{.irpc @var{symbol},@var{values}}@dots{}
3819 @cindex @code{irpc} directive
3820 Evaluate a sequence of statements assigning different values to @var{symbol}.
3821 The sequence of statements starts at the @code{.irpc} directive, and is
3822 terminated by an @code{.endr} directive. For each character in @var{value},
3823 @var{symbol} is set to the character, and the sequence of statements is
3824 assembled. If no @var{value} is listed, the sequence of statements is
3825 assembled once, with @var{symbol} set to the null string. To refer to
3826 @var{symbol} within the sequence of statements, use @var{\symbol}.
3828 For example, assembling
3836 is equivalent to assembling
3845 @section @code{.lcomm @var{symbol} , @var{length}}
3847 @cindex @code{lcomm} directive
3848 @cindex local common symbols
3849 @cindex symbols, local common
3850 Reserve @var{length} (an absolute expression) bytes for a local common
3851 denoted by @var{symbol}. The section and value of @var{symbol} are
3852 those of the new local common. The addresses are allocated in the bss
3853 section, so that at run-time the bytes start off zeroed. @var{Symbol}
3854 is not declared global (@pxref{Global,,@code{.global}}), so is normally
3855 not visible to @code{@value{LD}}.
3858 Some targets permit a third argument to be used with @code{.lcomm}. This
3859 argument specifies the desired alignment of the symbol in the bss section.
3863 The syntax for @code{.lcomm} differs slightly on the HPPA. The syntax is
3864 @samp{@var{symbol} .lcomm, @var{length}}; @var{symbol} is optional.
3868 @section @code{.lflags}
3870 @cindex @code{lflags} directive (ignored)
3871 @code{@value{AS}} accepts this directive, for compatibility with other
3872 assemblers, but ignores it.
3874 @ifclear no-line-dir
3876 @section @code{.line @var{line-number}}
3878 @cindex @code{line} directive
3882 @section @code{.ln @var{line-number}}
3884 @cindex @code{ln} directive
3886 @cindex logical line number
3888 Change the logical line number. @var{line-number} must be an absolute
3889 expression. The next line has that logical line number. Therefore any other
3890 statements on the current line (after a statement separator character) are
3891 reported as on logical line number @var{line-number} @minus{} 1. One day
3892 @code{@value{AS}} will no longer support this directive: it is recognized only
3893 for compatibility with existing assembler programs.
3897 @emph{Warning:} In the AMD29K configuration of @value{AS}, this command is
3898 not available; use the synonym @code{.ln} in that context.
3903 @ifclear no-line-dir
3904 Even though this is a directive associated with the @code{a.out} or
3905 @code{b.out} object-code formats, @code{@value{AS}} still recognizes it
3906 when producing COFF output, and treats @samp{.line} as though it
3907 were the COFF @samp{.ln} @emph{if} it is found outside a
3908 @code{.def}/@code{.endef} pair.
3910 Inside a @code{.def}, @samp{.line} is, instead, one of the directives
3911 used by compilers to generate auxiliary symbol information for
3916 @section @code{.linkonce [@var{type}]}
3918 @cindex @code{linkonce} directive
3919 @cindex common sections
3920 Mark the current section so that the linker only includes a single copy of it.
3921 This may be used to include the same section in several different object files,
3922 but ensure that the linker will only include it once in the final output file.
3923 The @code{.linkonce} pseudo-op must be used for each instance of the section.
3924 Duplicate sections are detected based on the section name, so it should be
3927 This directive is only supported by a few object file formats; as of this
3928 writing, the only object file format which supports it is the Portable
3929 Executable format used on Windows NT.
3931 The @var{type} argument is optional. If specified, it must be one of the
3932 following strings. For example:
3936 Not all types may be supported on all object file formats.
3940 Silently discard duplicate sections. This is the default.
3943 Warn if there are duplicate sections, but still keep only one copy.
3946 Warn if any of the duplicates have different sizes.
3949 Warn if any of the duplicates do not have exactly the same contents.
3953 @section @code{.ln @var{line-number}}
3955 @cindex @code{ln} directive
3956 @ifclear no-line-dir
3957 @samp{.ln} is a synonym for @samp{.line}.
3960 Tell @code{@value{AS}} to change the logical line number. @var{line-number}
3961 must be an absolute expression. The next line has that logical
3962 line number, so any other statements on the current line (after a
3963 statement separator character @code{;}) are reported as on logical
3964 line number @var{line-number} @minus{} 1.
3967 This directive is accepted, but ignored, when @code{@value{AS}} is
3968 configured for @code{b.out}; its effect is only associated with COFF
3974 @section @code{.mri @var{val}}
3976 @cindex @code{mri} directive
3977 @cindex MRI mode, temporarily
3978 If @var{val} is non-zero, this tells @code{@value{AS}} to enter MRI mode. If
3979 @var{val} is zero, this tells @code{@value{AS}} to exit MRI mode. This change
3980 affects code assembled until the next @code{.mri} directive, or until the end
3981 of the file. @xref{M, MRI mode, MRI mode}.
3984 @section @code{.list}
3986 @cindex @code{list} directive
3987 @cindex listing control, turning on
3988 Control (in conjunction with the @code{.nolist} directive) whether or
3989 not assembly listings are generated. These two directives maintain an
3990 internal counter (which is zero initially). @code{.list} increments the
3991 counter, and @code{.nolist} decrements it. Assembly listings are
3992 generated whenever the counter is greater than zero.
3994 By default, listings are disabled. When you enable them (with the
3995 @samp{-a} command line option; @pxref{Invoking,,Command-Line Options}),
3996 the initial value of the listing counter is one.
3999 @section @code{.long @var{expressions}}
4001 @cindex @code{long} directive
4002 @code{.long} is the same as @samp{.int}, @pxref{Int,,@code{.int}}.
4005 @c no one seems to know what this is for or whether this description is
4006 @c what it really ought to do
4008 @section @code{.lsym @var{symbol}, @var{expression}}
4010 @cindex @code{lsym} directive
4011 @cindex symbol, not referenced in assembly
4012 @code{.lsym} creates a new symbol named @var{symbol}, but does not put it in
4013 the hash table, ensuring it cannot be referenced by name during the
4014 rest of the assembly. This sets the attributes of the symbol to be
4015 the same as the expression value:
4017 @var{other} = @var{descriptor} = 0
4018 @var{type} = @r{(section of @var{expression})}
4019 @var{value} = @var{expression}
4022 The new symbol is not flagged as external.
4026 @section @code{.macro}
4029 The commands @code{.macro} and @code{.endm} allow you to define macros that
4030 generate assembly output. For example, this definition specifies a macro
4031 @code{sum} that puts a sequence of numbers into memory:
4034 .macro sum from=0, to=5
4043 With that definition, @samp{SUM 0,5} is equivalent to this assembly input:
4055 @item .macro @var{macname}
4056 @itemx .macro @var{macname} @var{macargs} @dots{}
4057 @cindex @code{macro} directive
4058 Begin the definition of a macro called @var{macname}. If your macro
4059 definition requires arguments, specify their names after the macro name,
4060 separated by commas or spaces. You can supply a default value for any
4061 macro argument by following the name with @samp{=@var{deflt}}. For
4062 example, these are all valid @code{.macro} statements:
4066 Begin the definition of a macro called @code{comm}, which takes no
4069 @item .macro plus1 p, p1
4070 @itemx .macro plus1 p p1
4071 Either statement begins the definition of a macro called @code{plus1},
4072 which takes two arguments; within the macro definition, write
4073 @samp{\p} or @samp{\p1} to evaluate the arguments.
4075 @item .macro reserve_str p1=0 p2
4076 Begin the definition of a macro called @code{reserve_str}, with two
4077 arguments. The first argument has a default value, but not the second.
4078 After the definition is complete, you can call the macro either as
4079 @samp{reserve_str @var{a},@var{b}} (with @samp{\p1} evaluating to
4080 @var{a} and @samp{\p2} evaluating to @var{b}), or as @samp{reserve_str
4081 ,@var{b}} (with @samp{\p1} evaluating as the default, in this case
4082 @samp{0}, and @samp{\p2} evaluating to @var{b}).
4085 When you call a macro, you can specify the argument values either by
4086 position, or by keyword. For example, @samp{sum 9,17} is equivalent to
4087 @samp{sum to=17, from=9}.
4090 @cindex @code{endm} directive
4091 Mark the end of a macro definition.
4094 @cindex @code{exitm} directive
4095 Exit early from the current macro definition.
4097 @cindex number of macros executed
4098 @cindex macros, count executed
4100 @code{@value{AS}} maintains a counter of how many macros it has
4101 executed in this pseudo-variable; you can copy that number to your
4102 output with @samp{\@@}, but @emph{only within a macro definition}.
4105 @item LOCAL @var{name} [ , @dots{} ]
4106 @emph{Warning: @code{LOCAL} is only available if you select ``alternate
4107 macro syntax'' with @samp{-a} or @samp{--alternate}.} @xref{Alternate,,
4108 Alternate macro syntax}.
4110 Generate a string replacement for each of the @var{name} arguments, and
4111 replace any instances of @var{name} in each macro expansion. The
4112 replacement string is unique in the assembly, and different for each
4113 separate macro expansion. @code{LOCAL} allows you to write macros that
4114 define symbols, without fear of conflict between separate macro expansions.
4119 @section @code{.nolist}
4121 @cindex @code{nolist} directive
4122 @cindex listing control, turning off
4123 Control (in conjunction with the @code{.list} directive) whether or
4124 not assembly listings are generated. These two directives maintain an
4125 internal counter (which is zero initially). @code{.list} increments the
4126 counter, and @code{.nolist} decrements it. Assembly listings are
4127 generated whenever the counter is greater than zero.
4130 @section @code{.octa @var{bignums}}
4132 @c FIXME: double size emitted for "octa" on i960, others? Or warn?
4133 @cindex @code{octa} directive
4134 @cindex integer, 16-byte
4135 @cindex sixteen byte integer
4136 This directive expects zero or more bignums, separated by commas. For each
4137 bignum, it emits a 16-byte integer.
4139 The term ``octa'' comes from contexts in which a ``word'' is two bytes;
4140 hence @emph{octa}-word for 16 bytes.
4143 @section @code{.org @var{new-lc} , @var{fill}}
4145 @cindex @code{org} directive
4146 @cindex location counter, advancing
4147 @cindex advancing location counter
4148 @cindex current address, advancing
4149 Advance the location counter of the current section to
4150 @var{new-lc}. @var{new-lc} is either an absolute expression or an
4151 expression with the same section as the current subsection. That is,
4152 you can't use @code{.org} to cross sections: if @var{new-lc} has the
4153 wrong section, the @code{.org} directive is ignored. To be compatible
4154 with former assemblers, if the section of @var{new-lc} is absolute,
4155 @code{@value{AS}} issues a warning, then pretends the section of @var{new-lc}
4156 is the same as the current subsection.
4158 @code{.org} may only increase the location counter, or leave it
4159 unchanged; you cannot use @code{.org} to move the location counter
4162 @c double negative used below "not undefined" because this is a specific
4163 @c reference to "undefined" (as SEG_UNKNOWN is called in this manual)
4164 @c section. doc@cygnus.com 18feb91
4165 Because @code{@value{AS}} tries to assemble programs in one pass, @var{new-lc}
4166 may not be undefined. If you really detest this restriction we eagerly await
4167 a chance to share your improved assembler.
4169 Beware that the origin is relative to the start of the section, not
4170 to the start of the subsection. This is compatible with other
4171 people's assemblers.
4173 When the location counter (of the current subsection) is advanced, the
4174 intervening bytes are filled with @var{fill} which should be an
4175 absolute expression. If the comma and @var{fill} are omitted,
4176 @var{fill} defaults to zero.
4179 @section @code{.p2align[wl] @var{abs-expr}, @var{abs-expr}, @var{abs-expr}}
4181 @cindex padding the location counter given a power of two
4182 @cindex @code{p2align} directive
4183 Pad the location counter (in the current subsection) to a particular
4184 storage boundary. The first expression (which must be absolute) is the
4185 number of low-order zero bits the location counter must have after
4186 advancement. For example @samp{.p2align 3} advances the location
4187 counter until it a multiple of 8. If the location counter is already a
4188 multiple of 8, no change is needed.
4190 The second expression (also absolute) gives the fill value to be stored in the
4191 padding bytes. It (and the comma) may be omitted. If it is omitted, the
4192 padding bytes are normally zero. However, on some systems, if the section is
4193 marked as containing code and the fill value is omitted, the space is filled
4194 with no-op instructions.
4196 The third expression is also absolute, and is also optional. If it is present,
4197 it is the maximum number of bytes that should be skipped by this alignment
4198 directive. If doing the alignment would require skipping more bytes than the
4199 specified maximum, then the alignment is not done at all. You can omit the
4200 fill value (the second argument) entirely by simply using two commas after the
4201 required alignment; this can be useful if you want the alignment to be filled
4202 with no-op instructions when appropriate.
4204 @cindex @code{p2alignw} directive
4205 @cindex @code{p2alignl} directive
4206 The @code{.p2alignw} and @code{.p2alignl} directives are variants of the
4207 @code{.p2align} directive. The @code{.p2alignw} directive treats the fill
4208 pattern as a two byte word value. The @code{.p2alignl} directives treats the
4209 fill pattern as a four byte longword value. For example, @code{.p2alignw
4210 2,0x368d} will align to a multiple of 4. If it skips two bytes, they will be
4211 filled in with the value 0x368d (the exact placement of the bytes depends upon
4212 the endianness of the processor). If it skips 1 or 3 bytes, the fill value is
4216 @section @code{.print @var{string}}
4218 @cindex @code{print} directive
4219 @code{@value{AS}} will print @var{string} on the standard output during
4220 assembly. You must put @var{string} in double quotes.
4223 @section @code{.psize @var{lines} , @var{columns}}
4225 @cindex @code{psize} directive
4226 @cindex listing control: paper size
4227 @cindex paper size, for listings
4228 Use this directive to declare the number of lines---and, optionally, the
4229 number of columns---to use for each page, when generating listings.
4231 If you do not use @code{.psize}, listings use a default line-count
4232 of 60. You may omit the comma and @var{columns} specification; the
4233 default width is 200 columns.
4235 @code{@value{AS}} generates formfeeds whenever the specified number of
4236 lines is exceeded (or whenever you explicitly request one, using
4239 If you specify @var{lines} as @code{0}, no formfeeds are generated save
4240 those explicitly specified with @code{.eject}.
4243 @section @code{.purgem @var{name}}
4245 @cindex @code{purgem} directive
4246 Undefine the macro @var{name}, so that later uses of the string will not be
4247 expanded. @xref{Macro}.
4250 @section @code{.quad @var{bignums}}
4252 @cindex @code{quad} directive
4253 @code{.quad} expects zero or more bignums, separated by commas. For
4254 each bignum, it emits
4256 an 8-byte integer. If the bignum won't fit in 8 bytes, it prints a
4257 warning message; and just takes the lowest order 8 bytes of the bignum.
4258 @cindex eight-byte integer
4259 @cindex integer, 8-byte
4261 The term ``quad'' comes from contexts in which a ``word'' is two bytes;
4262 hence @emph{quad}-word for 8 bytes.
4265 a 16-byte integer. If the bignum won't fit in 16 bytes, it prints a
4266 warning message; and just takes the lowest order 16 bytes of the bignum.
4267 @cindex sixteen-byte integer
4268 @cindex integer, 16-byte
4272 @section @code{.rept @var{count}}
4274 @cindex @code{rept} directive
4275 Repeat the sequence of lines between the @code{.rept} directive and the next
4276 @code{.endr} directive @var{count} times.
4278 For example, assembling
4286 is equivalent to assembling
4295 @section @code{.sbttl "@var{subheading}"}
4297 @cindex @code{sbttl} directive
4298 @cindex subtitles for listings
4299 @cindex listing control: subtitle
4300 Use @var{subheading} as the title (third line, immediately after the
4301 title line) when generating assembly listings.
4303 This directive affects subsequent pages, as well as the current page if
4304 it appears within ten lines of the top of a page.
4308 @section @code{.scl @var{class}}
4310 @cindex @code{scl} directive
4311 @cindex symbol storage class (COFF)
4312 @cindex COFF symbol storage class
4313 Set the storage-class value for a symbol. This directive may only be
4314 used inside a @code{.def}/@code{.endef} pair. Storage class may flag
4315 whether a symbol is static or external, or it may record further
4316 symbolic debugging information.
4319 The @samp{.scl} directive is primarily associated with COFF output; when
4320 configured to generate @code{b.out} output format, @code{@value{AS}}
4321 accepts this directive but ignores it.
4326 @section @code{.section @var{name}}
4328 @cindex @code{section} directive
4329 @cindex named section
4330 Use the @code{.section} directive to assemble the following code into a section
4333 This directive is only supported for targets that actually support arbitrarily
4334 named sections; on @code{a.out} targets, for example, it is not accepted, even
4335 with a standard @code{a.out} section name.
4338 For COFF targets, the @code{.section} directive is used in one of the following
4341 .section @var{name}[, "@var{flags}"]
4342 .section @var{name}[, @var{subsegment}]
4345 If the optional argument is quoted, it is taken as flags to use for the
4346 section. Each flag is a single character. The following flags are recognized:
4349 bss section (uninitialized data)
4351 section is not loaded
4361 shared section (meaningful for PE targets)
4364 If no flags are specified, the default flags depend upon the section name. If
4365 the section name is not recognized, the default will be for the section to be
4366 loaded and writable.
4368 If the optional argument to the @code{.section} directive is not quoted, it is
4369 taken as a subsegment number (@pxref{Sub-Sections}).
4373 For ELF targets, the @code{.section} directive is used like this:
4375 .section @var{name}[, "@var{flags}"[, @@@var{type}]]
4377 The optional @var{flags} argument is a quoted string which may contain any
4378 combintion of the following characters:
4381 section is allocatable
4385 section is executable
4388 The optional @var{type} argument may contain one of the following constants:
4391 section contains data
4393 section does not contain data (i.e., section only occupies space)
4396 If no flags are specified, the default flags depend upon the section name. If
4397 the section name is not recognized, the default will be for the section to have
4398 none of the above flags: it will not be allocated in memory, nor writable, nor
4399 executable. The section will contain data.
4401 For ELF targets, the assembler supports another type of @code{.section}
4402 directive for compatibility with the Solaris assembler:
4404 .section "@var{name}"[, @var{flags}...]
4406 Note that the section name is quoted. There may be a sequence of comma
4410 section is allocatable
4414 section is executable
4419 @section @code{.set @var{symbol}, @var{expression}}
4421 @cindex @code{set} directive
4422 @cindex symbol value, setting
4423 Set the value of @var{symbol} to @var{expression}. This
4424 changes @var{symbol}'s value and type to conform to
4425 @var{expression}. If @var{symbol} was flagged as external, it remains
4426 flagged (@pxref{Symbol Attributes}).
4428 You may @code{.set} a symbol many times in the same assembly.
4430 If you @code{.set} a global symbol, the value stored in the object
4431 file is the last value stored into it.
4434 The syntax for @code{set} on the HPPA is
4435 @samp{@var{symbol} .set @var{expression}}.
4439 @section @code{.short @var{expressions}}
4441 @cindex @code{short} directive
4443 @code{.short} is normally the same as @samp{.word}.
4444 @xref{Word,,@code{.word}}.
4446 In some configurations, however, @code{.short} and @code{.word} generate
4447 numbers of different lengths; @pxref{Machine Dependencies}.
4451 @code{.short} is the same as @samp{.word}. @xref{Word,,@code{.word}}.
4454 This expects zero or more @var{expressions}, and emits
4455 a 16 bit number for each.
4460 @section @code{.single @var{flonums}}
4462 @cindex @code{single} directive
4463 @cindex floating point numbers (single)
4464 This directive assembles zero or more flonums, separated by commas. It
4465 has the same effect as @code{.float}.
4467 The exact kind of floating point numbers emitted depends on how
4468 @code{@value{AS}} is configured. @xref{Machine Dependencies}.
4472 On the @value{TARGET} family, @code{.single} emits 32-bit floating point
4473 numbers in @sc{ieee} format.
4479 @section @code{.size}
4481 @cindex @code{size} directive
4482 This directive is generated by compilers to include auxiliary debugging
4483 information in the symbol table. It is only permitted inside
4484 @code{.def}/@code{.endef} pairs.
4487 @samp{.size} is only meaningful when generating COFF format output; when
4488 @code{@value{AS}} is generating @code{b.out}, it accepts this directive but
4494 @section @code{.sleb128 @var{expressions}}
4496 @cindex @code{sleb128} directive
4497 @var{sleb128} stands for ``signed little endian base 128.'' This is a
4498 compact, variable length representation of numbers used by the DWARF
4499 symbolic debugging format. @xref{Uleb128,@code{.uleb128}}.
4501 @ifclear no-space-dir
4503 @section @code{.skip @var{size} , @var{fill}}
4505 @cindex @code{skip} directive
4506 @cindex filling memory
4507 This directive emits @var{size} bytes, each of value @var{fill}. Both
4508 @var{size} and @var{fill} are absolute expressions. If the comma and
4509 @var{fill} are omitted, @var{fill} is assumed to be zero. This is the same as
4513 @section @code{.space @var{size} , @var{fill}}
4515 @cindex @code{space} directive
4516 @cindex filling memory
4517 This directive emits @var{size} bytes, each of value @var{fill}. Both
4518 @var{size} and @var{fill} are absolute expressions. If the comma
4519 and @var{fill} are omitted, @var{fill} is assumed to be zero. This is the same
4524 @emph{Warning:} @code{.space} has a completely different meaning for HPPA
4525 targets; use @code{.block} as a substitute. See @cite{HP9000 Series 800
4526 Assembly Language Reference Manual} (HP 92432-90001) for the meaning of the
4527 @code{.space} directive. @xref{HPPA Directives,,HPPA Assembler Directives},
4536 @section @code{.space}
4537 @cindex @code{space} directive
4539 On the AMD 29K, this directive is ignored; it is accepted for
4540 compatibility with other AMD 29K assemblers.
4543 @emph{Warning:} In most versions of the @sc{gnu} assembler, the directive
4544 @code{.space} has the effect of @code{.block} @xref{Machine Dependencies}.
4550 @section @code{.stabd, .stabn, .stabs}
4552 @cindex symbolic debuggers, information for
4553 @cindex @code{stab@var{x}} directives
4554 There are three directives that begin @samp{.stab}.
4555 All emit symbols (@pxref{Symbols}), for use by symbolic debuggers.
4556 The symbols are not entered in the @code{@value{AS}} hash table: they
4557 cannot be referenced elsewhere in the source file.
4558 Up to five fields are required:
4562 This is the symbol's name. It may contain any character except
4563 @samp{\000}, so is more general than ordinary symbol names. Some
4564 debuggers used to code arbitrarily complex structures into symbol names
4568 An absolute expression. The symbol's type is set to the low 8 bits of
4569 this expression. Any bit pattern is permitted, but @code{@value{LD}}
4570 and debuggers choke on silly bit patterns.
4573 An absolute expression. The symbol's ``other'' attribute is set to the
4574 low 8 bits of this expression.
4577 An absolute expression. The symbol's descriptor is set to the low 16
4578 bits of this expression.
4581 An absolute expression which becomes the symbol's value.
4584 If a warning is detected while reading a @code{.stabd}, @code{.stabn},
4585 or @code{.stabs} statement, the symbol has probably already been created;
4586 you get a half-formed symbol in your object file. This is
4587 compatible with earlier assemblers!
4590 @cindex @code{stabd} directive
4591 @item .stabd @var{type} , @var{other} , @var{desc}
4593 The ``name'' of the symbol generated is not even an empty string.
4594 It is a null pointer, for compatibility. Older assemblers used a
4595 null pointer so they didn't waste space in object files with empty
4598 The symbol's value is set to the location counter,
4599 relocatably. When your program is linked, the value of this symbol
4600 is the address of the location counter when the @code{.stabd} was
4603 @cindex @code{stabn} directive
4604 @item .stabn @var{type} , @var{other} , @var{desc} , @var{value}
4605 The name of the symbol is set to the empty string @code{""}.
4607 @cindex @code{stabs} directive
4608 @item .stabs @var{string} , @var{type} , @var{other} , @var{desc} , @var{value}
4609 All five fields are specified.
4615 @section @code{.string} "@var{str}"
4617 @cindex string, copying to object file
4618 @cindex @code{string} directive
4620 Copy the characters in @var{str} to the object file. You may specify more than
4621 one string to copy, separated by commas. Unless otherwise specified for a
4622 particular machine, the assembler marks the end of each string with a 0 byte.
4623 You can use any of the escape sequences described in @ref{Strings,,Strings}.
4626 @section @code{.struct @var{expression}}
4628 @cindex @code{struct} directive
4629 Switch to the absolute section, and set the section offset to @var{expression},
4630 which must be an absolute expression. You might use this as follows:
4639 This would define the symbol @code{field1} to have the value 0, the symbol
4640 @code{field2} to have the value 4, and the symbol @code{field3} to have the
4641 value 8. Assembly would be left in the absolute section, and you would need to
4642 use a @code{.section} directive of some sort to change to some other section
4643 before further assembly.
4647 @section @code{.symver}
4648 @cindex @code{symver} directive
4649 @cindex symbol versioning
4650 @cindex versions of symbols
4651 Use the @code{.symver} directive to bind symbols to specific version nodes
4652 within a source file. This is only supported on ELF platforms, and is
4653 typically used when assembling files to be linked into a shared library.
4654 There are cases where it may make sense to use this in objects to be bound
4655 into an application itself so as to override a versioned symbol from a
4658 For ELF targets, the @code{.symver} directive is used like this:
4660 .symver @var{name}, @var{name2@@nodename}
4662 In this case, the symbol @var{name} must exist and be defined within the file
4663 being assembled. The @code{.versym} directive effectively creates a symbol
4664 alias with the name @var{name2@@nodename}, and in fact the main reason that we
4665 just don't try and create a regular alias is that the @var{@@} character isn't
4666 permitted in symbol names. The @var{name2} part of the name is the actual name
4667 of the symbol by which it will be externally referenced. The name @var{name}
4668 itself is merely a name of convenience that is used so that it is possible to
4669 have definitions for multiple versions of a function within a single source
4670 file, and so that the compiler can unambiguously know which version of a
4671 function is being mentioned. The @var{nodename} portion of the alias should be
4672 the name of a node specified in the version script supplied to the linker when
4673 building a shared library. If you are attempting to override a versioned
4674 symbol from a shared library, then @var{nodename} should correspond to the
4675 nodename of the symbol you are trying to override.
4680 @section @code{.tag @var{structname}}
4682 @cindex COFF structure debugging
4683 @cindex structure debugging, COFF
4684 @cindex @code{tag} directive
4685 This directive is generated by compilers to include auxiliary debugging
4686 information in the symbol table. It is only permitted inside
4687 @code{.def}/@code{.endef} pairs. Tags are used to link structure
4688 definitions in the symbol table with instances of those structures.
4691 @samp{.tag} is only used when generating COFF format output; when
4692 @code{@value{AS}} is generating @code{b.out}, it accepts this directive but
4698 @section @code{.text @var{subsection}}
4700 @cindex @code{text} directive
4701 Tells @code{@value{AS}} to assemble the following statements onto the end of
4702 the text subsection numbered @var{subsection}, which is an absolute
4703 expression. If @var{subsection} is omitted, subsection number zero
4707 @section @code{.title "@var{heading}"}
4709 @cindex @code{title} directive
4710 @cindex listing control: title line
4711 Use @var{heading} as the title (second line, immediately after the
4712 source file name and pagenumber) when generating assembly listings.
4714 This directive affects subsequent pages, as well as the current page if
4715 it appears within ten lines of the top of a page.
4719 @section @code{.type @var{int}}
4721 @cindex COFF symbol type
4722 @cindex symbol type, COFF
4723 @cindex @code{type} directive
4724 This directive, permitted only within @code{.def}/@code{.endef} pairs,
4725 records the integer @var{int} as the type attribute of a symbol table entry.
4728 @samp{.type} is associated only with COFF format output; when
4729 @code{@value{AS}} is configured for @code{b.out} output, it accepts this
4730 directive but ignores it.
4736 @section @code{.val @var{addr}}
4738 @cindex @code{val} directive
4739 @cindex COFF value attribute
4740 @cindex value attribute, COFF
4741 This directive, permitted only within @code{.def}/@code{.endef} pairs,
4742 records the address @var{addr} as the value attribute of a symbol table
4746 @samp{.val} is used only for COFF output; when @code{@value{AS}} is
4747 configured for @code{b.out}, it accepts this directive but ignores it.
4752 @section @code{.uleb128 @var{expressions}}
4754 @cindex @code{uleb128} directive
4755 @var{uleb128} stands for ``unsigned little endian base 128.'' This is a
4756 compact, variable length representation of numbers used by the DWARF
4757 symbolic debugging format. @xref{Sleb128,@code{.sleb128}}.
4760 @section @code{.word @var{expressions}}
4762 @cindex @code{word} directive
4763 This directive expects zero or more @var{expressions}, of any section,
4764 separated by commas.
4767 For each expression, @code{@value{AS}} emits a 32-bit number.
4770 For each expression, @code{@value{AS}} emits a 16-bit number.
4775 The size of the number emitted, and its byte order,
4776 depend on what target computer the assembly is for.
4779 @c on amd29k, i960, sparc the "special treatment to support compilers" doesn't
4780 @c happen---32-bit addressability, period; no long/short jumps.
4781 @ifset DIFF-TBL-KLUGE
4782 @cindex difference tables altered
4783 @cindex altered difference tables
4785 @emph{Warning: Special Treatment to support Compilers}
4789 Machines with a 32-bit address space, but that do less than 32-bit
4790 addressing, require the following special treatment. If the machine of
4791 interest to you does 32-bit addressing (or doesn't require it;
4792 @pxref{Machine Dependencies}), you can ignore this issue.
4795 In order to assemble compiler output into something that works,
4796 @code{@value{AS}} occasionlly does strange things to @samp{.word} directives.
4797 Directives of the form @samp{.word sym1-sym2} are often emitted by
4798 compilers as part of jump tables. Therefore, when @code{@value{AS}} assembles a
4799 directive of the form @samp{.word sym1-sym2}, and the difference between
4800 @code{sym1} and @code{sym2} does not fit in 16 bits, @code{@value{AS}}
4801 creates a @dfn{secondary jump table}, immediately before the next label.
4802 This secondary jump table is preceded by a short-jump to the
4803 first byte after the secondary table. This short-jump prevents the flow
4804 of control from accidentally falling into the new table. Inside the
4805 table is a long-jump to @code{sym2}. The original @samp{.word}
4806 contains @code{sym1} minus the address of the long-jump to
4809 If there were several occurrences of @samp{.word sym1-sym2} before the
4810 secondary jump table, all of them are adjusted. If there was a
4811 @samp{.word sym3-sym4}, that also did not fit in sixteen bits, a
4812 long-jump to @code{sym4} is included in the secondary jump table,
4813 and the @code{.word} directives are adjusted to contain @code{sym3}
4814 minus the address of the long-jump to @code{sym4}; and so on, for as many
4815 entries in the original jump table as necessary.
4818 @emph{This feature may be disabled by compiling @code{@value{AS}} with the
4819 @samp{-DWORKING_DOT_WORD} option.} This feature is likely to confuse
4820 assembly language programmers.
4823 @c end DIFF-TBL-KLUGE
4826 @section Deprecated Directives
4828 @cindex deprecated directives
4829 @cindex obsolescent directives
4830 One day these directives won't work.
4831 They are included for compatibility with older assemblers.
4838 @node Machine Dependencies
4839 @chapter Machine Dependent Features
4841 @cindex machine dependencies
4842 The machine instruction sets are (almost by definition) different on
4843 each machine where @code{@value{AS}} runs. Floating point representations
4844 vary as well, and @code{@value{AS}} often supports a few additional
4845 directives or command-line options for compatibility with other
4846 assemblers on a particular platform. Finally, some versions of
4847 @code{@value{AS}} support special pseudo-instructions for branch
4850 This chapter discusses most of these differences, though it does not
4851 include details on any machine's instruction set. For details on that
4852 subject, see the hardware manufacturer's manual.
4856 * AMD29K-Dependent:: AMD 29K Dependent Features
4859 * ARC-Dependent:: ARC Dependent Features
4862 * ARM-Dependent:: ARM Dependent Features
4865 * D10V-Dependent:: D10V Dependent Features
4868 * D30V-Dependent:: D30V Dependent Features
4871 * H8/300-Dependent:: Hitachi H8/300 Dependent Features
4874 * H8/500-Dependent:: Hitachi H8/500 Dependent Features
4877 * HPPA-Dependent:: HPPA Dependent Features
4880 * i386-Dependent:: Intel 80386 Dependent Features
4883 * i960-Dependent:: Intel 80960 Dependent Features
4886 * M68K-Dependent:: M680x0 Dependent Features
4889 * MIPS-Dependent:: MIPS Dependent Features
4892 * SH-Dependent:: Hitachi SH Dependent Features
4895 * PJ-Dependent:: picoJava Dependent Features
4898 * Sparc-Dependent:: SPARC Dependent Features
4901 * V850-Dependent:: V850 Dependent Features
4904 * Z8000-Dependent:: Z8000 Dependent Features
4907 * Vax-Dependent:: VAX Dependent Features
4914 @c The following major nodes are *sections* in the GENERIC version, *chapters*
4915 @c in single-cpu versions. This is mainly achieved by @lowersections. There is a
4916 @c peculiarity: to preserve cross-references, there must be a node called
4917 @c "Machine Dependencies". Hence the conditional nodenames in each
4918 @c major node below. Node defaulting in makeinfo requires adjacency of
4919 @c node and sectioning commands; hence the repetition of @chapter BLAH
4920 @c in both conditional blocks.
4926 @chapter ARC Dependent Features
4929 @node Machine Dependencies
4930 @chapter ARC Dependent Features
4935 * ARC-Opts:: Options
4936 * ARC-Float:: Floating Point
4937 * ARC-Directives:: Sparc Machine Directives
4943 @cindex options for ARC
4945 @cindex architectures, ARC
4946 @cindex ARC architectures
4947 The ARC chip family includes several successive levels (or other
4948 variants) of chip, using the same core instruction set, but including
4949 a few additional instructions at each level.
4951 By default, @code{@value{AS}} assumes the core instruction set (ARC
4952 base). The @code{.cpu} pseudo-op is intended to be used to select
4956 @cindex @code{-mbig-endian} option (ARC)
4957 @cindex @code{-mlittle-endian} option (ARC)
4958 @cindex ARC big-endian output
4959 @cindex ARC little-endian output
4960 @cindex big-endian output, ARC
4961 @cindex little-endian output, ARC
4963 @itemx -mlittle-endian
4964 Any @sc{arc} configuration of @code{@value{AS}} can select big-endian or
4965 little-endian output at run time (unlike most other @sc{gnu} development
4966 tools, which must be configured for one or the other). Use
4967 @samp{-mbig-endian} to select big-endian output, and @samp{-mlittle-endian}
4972 @section Floating Point
4974 @cindex floating point, ARC (@sc{ieee})
4975 @cindex ARC floating point (@sc{ieee})
4976 The ARC cpu family currently does not have hardware floating point
4977 support. Software floating point support is provided by @code{GCC}
4978 and uses @sc{ieee} floating-point numbers.
4980 @node ARC-Directives
4981 @section ARC Machine Directives
4983 @cindex ARC machine directives
4984 @cindex machine directives, ARC
4985 The ARC version of @code{@value{AS}} supports the following additional
4990 @cindex @code{cpu} directive, SPARC
4991 This must be followed by the desired cpu.
4992 The ARC is intended to be customizable, @code{.cpu} is used to
4993 select the desired variant [though currently there are none].
5000 @include c-a29k.texi
5009 @node Machine Dependencies
5010 @chapter Machine Dependent Features
5012 The machine instruction sets are different on each Hitachi chip family,
5013 and there are also some syntax differences among the families. This
5014 chapter describes the specific @code{@value{AS}} features for each
5018 * H8/300-Dependent:: Hitachi H8/300 Dependent Features
5019 * H8/500-Dependent:: Hitachi H8/500 Dependent Features
5020 * SH-Dependent:: Hitachi SH Dependent Features
5027 @include c-d10v.texi
5031 @include c-d30v.texi
5035 @include c-h8300.texi
5039 @include c-h8500.texi
5043 @include c-hppa.texi
5047 @include c-i386.texi
5051 @include c-i960.texi
5056 @include c-m68k.texi
5060 @include c-mips.texi
5064 @include c-ns32k.texi
5076 @include c-sparc.texi
5088 @include c-v850.texi
5092 @c reverse effect of @down at top of generic Machine-Dep chapter
5096 @node Reporting Bugs
5097 @chapter Reporting Bugs
5098 @cindex bugs in assembler
5099 @cindex reporting bugs in assembler
5101 Your bug reports play an essential role in making @code{@value{AS}} reliable.
5103 Reporting a bug may help you by bringing a solution to your problem, or it may
5104 not. But in any case the principal function of a bug report is to help the
5105 entire community by making the next version of @code{@value{AS}} work better.
5106 Bug reports are your contribution to the maintenance of @code{@value{AS}}.
5108 In order for a bug report to serve its purpose, you must include the
5109 information that enables us to fix the bug.
5112 * Bug Criteria:: Have you found a bug?
5113 * Bug Reporting:: How to report bugs
5117 @section Have you found a bug?
5118 @cindex bug criteria
5120 If you are not sure whether you have found a bug, here are some guidelines:
5123 @cindex fatal signal
5124 @cindex assembler crash
5125 @cindex crash of assembler
5127 If the assembler gets a fatal signal, for any input whatever, that is a
5128 @code{@value{AS}} bug. Reliable assemblers never crash.
5130 @cindex error on valid input
5132 If @code{@value{AS}} produces an error message for valid input, that is a bug.
5134 @cindex invalid input
5136 If @code{@value{AS}} does not produce an error message for invalid input, that
5137 is a bug. However, you should note that your idea of ``invalid input'' might
5138 be our idea of ``an extension'' or ``support for traditional practice''.
5141 If you are an experienced user of assemblers, your suggestions for improvement
5142 of @code{@value{AS}} are welcome in any case.
5146 @section How to report bugs
5148 @cindex assembler bugs, reporting
5150 A number of companies and individuals offer support for @sc{gnu} products. If
5151 you obtained @code{@value{AS}} from a support organization, we recommend you
5152 contact that organization first.
5154 You can find contact information for many support companies and
5155 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
5158 In any event, we also recommend that you send bug reports for @code{@value{AS}}
5159 to @samp{bug-gnu-utils@@gnu.org}.
5161 The fundamental principle of reporting bugs usefully is this:
5162 @strong{report all the facts}. If you are not sure whether to state a
5163 fact or leave it out, state it!
5165 Often people omit facts because they think they know what causes the problem
5166 and assume that some details do not matter. Thus, you might assume that the
5167 name of a symbol you use in an example does not matter. Well, probably it does
5168 not, but one cannot be sure. Perhaps the bug is a stray memory reference which
5169 happens to fetch from the location where that name is stored in memory;
5170 perhaps, if the name were different, the contents of that location would fool
5171 the assembler into doing the right thing despite the bug. Play it safe and
5172 give a specific, complete example. That is the easiest thing for you to do,
5173 and the most helpful.
5175 Keep in mind that the purpose of a bug report is to enable us to fix the bug if
5176 it is new to us. Therefore, always write your bug reports on the assumption
5177 that the bug has not been reported previously.
5179 Sometimes people give a few sketchy facts and ask, ``Does this ring a
5180 bell?'' Those bug reports are useless, and we urge everyone to
5181 @emph{refuse to respond to them} except to chide the sender to report
5184 To enable us to fix the bug, you should include all these things:
5188 The version of @code{@value{AS}}. @code{@value{AS}} announces it if you start
5189 it with the @samp{--version} argument.
5191 Without this, we will not know whether there is any point in looking for
5192 the bug in the current version of @code{@value{AS}}.
5195 Any patches you may have applied to the @code{@value{AS}} source.
5198 The type of machine you are using, and the operating system name and
5202 What compiler (and its version) was used to compile @code{@value{AS}}---e.g.
5206 The command arguments you gave the assembler to assemble your example and
5207 observe the bug. To guarantee you will not omit something important, list them
5208 all. A copy of the Makefile (or the output from make) is sufficient.
5210 If we were to try to guess the arguments, we would probably guess wrong
5211 and then we might not encounter the bug.
5214 A complete input file that will reproduce the bug. If the bug is observed when
5215 the assembler is invoked via a compiler, send the assembler source, not the
5216 high level language source. Most compilers will produce the assembler source
5217 when run with the @samp{-S} option. If you are using @code{@value{GCC}}, use
5218 the options @samp{-v --save-temps}; this will save the assembler source in a
5219 file with an extension of @file{.s}, and also show you exactly how
5220 @code{@value{AS}} is being run.
5223 A description of what behavior you observe that you believe is
5224 incorrect. For example, ``It gets a fatal signal.''
5226 Of course, if the bug is that @code{@value{AS}} gets a fatal signal, then we
5227 will certainly notice it. But if the bug is incorrect output, we might not
5228 notice unless it is glaringly wrong. You might as well not give us a chance to
5231 Even if the problem you experience is a fatal signal, you should still say so
5232 explicitly. Suppose something strange is going on, such as, your copy of
5233 @code{@value{AS}} is out of synch, or you have encountered a bug in the C
5234 library on your system. (This has happened!) Your copy might crash and ours
5235 would not. If you told us to expect a crash, then when ours fails to crash, we
5236 would know that the bug was not happening for us. If you had not told us to
5237 expect a crash, then we would not be able to draw any conclusion from our
5241 If you wish to suggest changes to the @code{@value{AS}} source, send us context
5242 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or @samp{-p}
5243 option. Always send diffs from the old file to the new file. If you even
5244 discuss something in the @code{@value{AS}} source, refer to it by context, not
5247 The line numbers in our development sources will not match those in your
5248 sources. Your line numbers would convey no useful information to us.
5251 Here are some things that are not necessary:
5255 A description of the envelope of the bug.
5257 Often people who encounter a bug spend a lot of time investigating
5258 which changes to the input file will make the bug go away and which
5259 changes will not affect it.
5261 This is often time consuming and not very useful, because the way we
5262 will find the bug is by running a single example under the debugger
5263 with breakpoints, not by pure deduction from a series of examples.
5264 We recommend that you save your time for something else.
5266 Of course, if you can find a simpler example to report @emph{instead}
5267 of the original one, that is a convenience for us. Errors in the
5268 output will be easier to spot, running under the debugger will take
5269 less time, and so on.
5271 However, simplification is not vital; if you do not want to do this,
5272 report the bug anyway and send us the entire test case you used.
5275 A patch for the bug.
5277 A patch for the bug does help us if it is a good one. But do not omit
5278 the necessary information, such as the test case, on the assumption that
5279 a patch is all we need. We might see problems with your patch and decide
5280 to fix the problem another way, or we might not understand it at all.
5282 Sometimes with a program as complicated as @code{@value{AS}} it is very hard to
5283 construct an example that will make the program follow a certain path through
5284 the code. If you do not send us the example, we will not be able to construct
5285 one, so we will not be able to verify that the bug is fixed.
5287 And if we cannot understand what bug you are trying to fix, or why your
5288 patch should be an improvement, we will not install it. A test case will
5289 help us to understand.
5292 A guess about what the bug is or what it depends on.
5294 Such guesses are usually wrong. Even we cannot guess right about such
5295 things without first using the debugger to find the facts.
5298 @node Acknowledgements
5299 @chapter Acknowledgements
5301 If you have contributed to @code{@value{AS}} and your name isn't listed here,
5302 it is not meant as a slight. We just don't know about it. Send mail to the
5303 maintainer, and we'll correct the situation. Currently
5305 the maintainer is Ken Raeburn (email address @code{raeburn@@cygnus.com}).
5307 Dean Elsner wrote the original @sc{gnu} assembler for the VAX.@footnote{Any
5310 Jay Fenlason maintained GAS for a while, adding support for GDB-specific debug
5311 information and the 68k series machines, most of the preprocessing pass, and
5312 extensive changes in @file{messages.c}, @file{input-file.c}, @file{write.c}.
5314 K. Richard Pixley maintained GAS for a while, adding various enhancements and
5315 many bug fixes, including merging support for several processors, breaking GAS
5316 up to handle multiple object file format back ends (including heavy rewrite,
5317 testing, an integration of the coff and b.out back ends), adding configuration
5318 including heavy testing and verification of cross assemblers and file splits
5319 and renaming, converted GAS to strictly ANSI C including full prototypes, added
5320 support for m680[34]0 and cpu32, did considerable work on i960 including a COFF
5321 port (including considerable amounts of reverse engineering), a SPARC opcode
5322 file rewrite, DECstation, rs6000, and hp300hpux host ports, updated ``know''
5323 assertions and made them work, much other reorganization, cleanup, and lint.
5325 Ken Raeburn wrote the high-level BFD interface code to replace most of the code
5326 in format-specific I/O modules.
5328 The original VMS support was contributed by David L. Kashtan. Eric Youngdale
5329 has done much work with it since.
5331 The Intel 80386 machine description was written by Eliot Dresselhaus.
5333 Minh Tran-Le at IntelliCorp contributed some AIX 386 support.
5335 The Motorola 88k machine description was contributed by Devon Bowen of Buffalo
5336 University and Torbjorn Granlund of the Swedish Institute of Computer Science.
5338 Keith Knowles at the Open Software Foundation wrote the original MIPS back end
5339 (@file{tc-mips.c}, @file{tc-mips.h}), and contributed Rose format support
5340 (which hasn't been merged in yet). Ralph Campbell worked with the MIPS code to
5341 support a.out format.
5343 Support for the Zilog Z8k and Hitachi H8/300 and H8/500 processors (tc-z8k,
5344 tc-h8300, tc-h8500), and IEEE 695 object file format (obj-ieee), was written by
5345 Steve Chamberlain of Cygnus Support. Steve also modified the COFF back end to
5346 use BFD for some low-level operations, for use with the H8/300 and AMD 29k
5349 John Gilmore built the AMD 29000 support, added @code{.include} support, and
5350 simplified the configuration of which versions accept which directives. He
5351 updated the 68k machine description so that Motorola's opcodes always produced
5352 fixed-size instructions (e.g. @code{jsr}), while synthetic instructions
5353 remained shrinkable (@code{jbsr}). John fixed many bugs, including true tested
5354 cross-compilation support, and one bug in relaxation that took a week and
5355 required the proverbial one-bit fix.
5357 Ian Lance Taylor of Cygnus Support merged the Motorola and MIT syntax for the
5358 68k, completed support for some COFF targets (68k, i386 SVR3, and SCO Unix),
5359 added support for MIPS ECOFF and ELF targets, wrote the initial RS/6000 and
5360 PowerPC assembler, and made a few other minor patches.
5362 Steve Chamberlain made @code{@value{AS}} able to generate listings.
5364 Hewlett-Packard contributed support for the HP9000/300.
5366 Jeff Law wrote GAS and BFD support for the native HPPA object format (SOM)
5367 along with a fairly extensive HPPA testsuite (for both SOM and ELF object
5368 formats). This work was supported by both the Center for Software Science at
5369 the University of Utah and Cygnus Support.
5371 Support for ELF format files has been worked on by Mark Eichin of Cygnus
5372 Support (original, incomplete implementation for SPARC), Pete Hoogenboom and
5373 Jeff Law at the University of Utah (HPPA mainly), Michael Meissner of the Open
5374 Software Foundation (i386 mainly), and Ken Raeburn of Cygnus Support (sparc,
5375 and some initial 64-bit support).
5377 Richard Henderson rewrote the Alpha assembler. Klaus Kaempf wrote GAS and BFD
5378 support for openVMS/Alpha.
5380 Several engineers at Cygnus Support have also provided many small bug fixes and
5381 configuration enhancements.
5383 Many others have contributed large or small bugfixes and enhancements. If
5384 you have contributed significant work and are not mentioned on this list, and
5385 want to be, let us know. Some of the history has been lost; we are not
5386 intentionally leaving anyone out.