Add "?" pseudo-flag to ELF .section/.pushsection directives.
[deliverable/binutils-gdb.git] / gas / doc / as.texinfo
CommitLineData
252b5132 1\input texinfo @c -*-Texinfo-*-
f7e42eb4 2@c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
40b36596 3@c 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
252b5132
RH
4@c Free Software Foundation, Inc.
5@c UPDATE!! On future updates--
6@c (1) check for new machine-dep cmdline options in
7@c md_parse_option definitions in config/tc-*.c
8@c (2) for platform-specific directives, examine md_pseudo_op
9@c in config/tc-*.c
10@c (3) for object-format specific directives, examine obj_pseudo_op
01642c12 11@c in config/obj-*.c
252b5132
RH
12@c (4) portable directives in potable[] in read.c
13@c %**start of header
14@setfilename as.info
15@c ---config---
a4fb0134
SC
16@macro gcctabopt{body}
17@code{\body\}
18@end macro
252b5132
RH
19@c defaults, config file may override:
20@set have-stabs
21@c ---
4a4c4a1d
MR
22@c man begin NAME
23@c ---
252b5132 24@include asconfig.texi
c428fa83 25@include bfdver.texi
252b5132 26@c ---
0285c67d 27@c man end
4a4c4a1d 28@c ---
252b5132 29@c common OR combinations of conditions
c1253627
NC
30@ifset COFF
31@set COFF-ELF
32@end ifset
33@ifset ELF
34@set COFF-ELF
35@end ifset
252b5132
RH
36@ifset AOUT
37@set aout-bout
38@end ifset
39@ifset ARM/Thumb
40@set ARM
41@end ifset
9982501a
JZ
42@ifset Blackfin
43@set Blackfin
44@end ifset
252b5132
RH
45@ifset BOUT
46@set aout-bout
47@end ifset
48@ifset H8/300
49@set H8
50@end ifset
252b5132
RH
51@ifset SH
52@set H8
53@end ifset
54@ifset HPPA
55@set abnormal-separator
56@end ifset
57@c ------------
58@ifset GENERIC
59@settitle Using @value{AS}
60@end ifset
61@ifclear GENERIC
62@settitle Using @value{AS} (@value{TARGET})
63@end ifclear
64@setchapternewpage odd
65@c %**end of header
66
67@c @smallbook
68@c @set SMALL
69@c WARE! Some of the machine-dependent sections contain tables of machine
70@c instructions. Except in multi-column format, these tables look silly.
71@c Unfortunately, Texinfo doesn't have a general-purpose multi-col format, so
72@c the multi-col format is faked within @example sections.
01642c12 73@c
252b5132
RH
74@c Again unfortunately, the natural size that fits on a page, for these tables,
75@c is different depending on whether or not smallbook is turned on.
76@c This matters, because of order: text flow switches columns at each page
77@c break.
01642c12 78@c
252b5132
RH
79@c The format faked in this source works reasonably well for smallbook,
80@c not well for the default large-page format. This manual expects that if you
81@c turn on @smallbook, you will also uncomment the "@set SMALL" to enable the
82@c tables in question. You can turn on one without the other at your
01642c12 83@c discretion, of course.
252b5132
RH
84@ifinfo
85@set SMALL
86@c the insn tables look just as silly in info files regardless of smallbook,
87@c might as well show 'em anyways.
88@end ifinfo
89
90@ifinfo
91@format
92START-INFO-DIR-ENTRY
93* As: (as). The GNU assembler.
59455fb1 94* Gas: (as). The GNU assembler.
252b5132
RH
95END-INFO-DIR-ENTRY
96@end format
97@end ifinfo
98
99@finalout
100@syncodeindex ky cp
101
0e9517a9 102@copying
252b5132
RH
103This file documents the GNU Assembler "@value{AS}".
104
0285c67d 105@c man begin COPYRIGHT
9fbcbd81
NC
106Copyright @copyright{} 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
1072000, 2001, 2002, 2006, 2007, 2008, 2009, 2010 Free Software Foundation, Inc.
252b5132 108
0285c67d 109Permission is granted to copy, distribute and/or modify this document
793c5807 110under the terms of the GNU Free Documentation License, Version 1.3
0285c67d
NC
111or any later version published by the Free Software Foundation;
112with no Invariant Sections, with no Front-Cover Texts, and with no
113Back-Cover Texts. A copy of the license is included in the
c1253627 114section entitled ``GNU Free Documentation License''.
0285c67d
NC
115
116@c man end
0e9517a9 117@end copying
252b5132
RH
118
119@titlepage
120@title Using @value{AS}
121@subtitle The @sc{gnu} Assembler
122@ifclear GENERIC
123@subtitle for the @value{TARGET} family
124@end ifclear
e49e529d
JM
125@ifset VERSION_PACKAGE
126@sp 1
127@subtitle @value{VERSION_PACKAGE}
128@end ifset
252b5132
RH
129@sp 1
130@subtitle Version @value{VERSION}
131@sp 1
132@sp 13
b45619c0 133The Free Software Foundation Inc.@: thanks The Nice Computer
252b5132 134Company of Australia for loaning Dean Elsner to write the
a4fb0134 135first (Vax) version of @command{as} for Project @sc{gnu}.
252b5132
RH
136The proprietors, management and staff of TNCCA thank FSF for
137distracting the boss while they got some work
138done.
139@sp 3
140@author Dean Elsner, Jay Fenlason & friends
141@page
142@tex
143{\parskip=0pt
144\hfill {\it Using {\tt @value{AS}}}\par
145\hfill Edited by Cygnus Support\par
146}
147%"boxit" macro for figures:
148%Modified from Knuth's ``boxit'' macro from TeXbook (answer to exercise 21.3)
149\gdef\boxit#1#2{\vbox{\hrule\hbox{\vrule\kern3pt
150 \vbox{\parindent=0pt\parskip=0pt\hsize=#1\kern3pt\strut\hfil
151#2\hfil\strut\kern3pt}\kern3pt\vrule}\hrule}}%box with visible outline
152\gdef\ibox#1#2{\hbox to #1{#2\hfil}\kern8pt}% invisible box
153@end tex
154
155@vskip 0pt plus 1filll
9fbcbd81
NC
156Copyright @copyright{} 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
1572000, 2001, 2002, 2006, 2007, 2008, 2009, 2010 Free Software Foundation, Inc.
252b5132 158
cf055d54 159 Permission is granted to copy, distribute and/or modify this document
793c5807 160 under the terms of the GNU Free Documentation License, Version 1.3
cf055d54
NC
161 or any later version published by the Free Software Foundation;
162 with no Invariant Sections, with no Front-Cover Texts, and with no
163 Back-Cover Texts. A copy of the license is included in the
c1253627 164 section entitled ``GNU Free Documentation License''.
252b5132 165
252b5132 166@end titlepage
4ecceb71 167@contents
252b5132 168
2e64b665 169@ifnottex
252b5132
RH
170@node Top
171@top Using @value{AS}
172
e49e529d
JM
173This file is a user guide to the @sc{gnu} assembler @command{@value{AS}}
174@ifset VERSION_PACKAGE
175@value{VERSION_PACKAGE}
176@end ifset
177version @value{VERSION}.
252b5132 178@ifclear GENERIC
a4fb0134 179This version of the file describes @command{@value{AS}} configured to generate
252b5132
RH
180code for @value{TARGET} architectures.
181@end ifclear
cf055d54
NC
182
183This document is distributed under the terms of the GNU Free
184Documentation License. A copy of the license is included in the
c1253627 185section entitled ``GNU Free Documentation License''.
cf055d54 186
252b5132
RH
187@menu
188* Overview:: Overview
189* Invoking:: Command-Line Options
190* Syntax:: Syntax
191* Sections:: Sections and Relocation
192* Symbols:: Symbols
193* Expressions:: Expressions
194* Pseudo Ops:: Assembler Directives
3a99f02f
DJ
195@ifset ELF
196* Object Attributes:: Object Attributes
197@end ifset
252b5132
RH
198* Machine Dependencies:: Machine Dependent Features
199* Reporting Bugs:: Reporting Bugs
200* Acknowledgements:: Who Did What
cf055d54 201* GNU Free Documentation License:: GNU Free Documentation License
28c9d252 202* AS Index:: AS Index
252b5132 203@end menu
2e64b665 204@end ifnottex
252b5132
RH
205
206@node Overview
207@chapter Overview
208@iftex
a4fb0134 209This manual is a user guide to the @sc{gnu} assembler @command{@value{AS}}.
252b5132 210@ifclear GENERIC
a4fb0134 211This version of the manual describes @command{@value{AS}} configured to generate
252b5132
RH
212code for @value{TARGET} architectures.
213@end ifclear
214@end iftex
215
216@cindex invocation summary
217@cindex option summary
218@cindex summary of options
a4fb0134 219Here is a brief summary of how to invoke @command{@value{AS}}. For details,
96e9638b 220see @ref{Invoking,,Command-Line Options}.
252b5132 221
0285c67d
NC
222@c man title AS the portable GNU assembler.
223
a4fb0134 224@ignore
0285c67d
NC
225@c man begin SEEALSO
226gcc(1), ld(1), and the Info entries for @file{binutils} and @file{ld}.
227@c man end
a4fb0134 228@end ignore
0285c67d 229
252b5132
RH
230@c We don't use deffn and friends for the following because they seem
231@c to be limited to one line for the header.
232@smallexample
0285c67d 233@c man begin SYNOPSIS
83f10cb2 234@value{AS} [@b{-a}[@b{cdghlns}][=@var{file}]] [@b{--alternate}] [@b{-D}]
955974c6 235 [@b{--compress-debug-sections}] [@b{--nocompress-debug-sections}]
3d6b762c 236 [@b{--debug-prefix-map} @var{old}=@var{new}]
4bdd3565
NC
237 [@b{--defsym} @var{sym}=@var{val}] [@b{-f}] [@b{-g}] [@b{--gstabs}]
238 [@b{--gstabs+}] [@b{--gdwarf-2}] [@b{--help}] [@b{-I} @var{dir}] [@b{-J}]
239 [@b{-K}] [@b{-L}] [@b{--listing-lhs-width}=@var{NUM}]
240 [@b{--listing-lhs-width2}=@var{NUM}] [@b{--listing-rhs-width}=@var{NUM}]
241 [@b{--listing-cont-lines}=@var{NUM}] [@b{--keep-locals}] [@b{-o}
242 @var{objfile}] [@b{-R}] [@b{--reduce-memory-overheads}] [@b{--statistics}]
243 [@b{-v}] [@b{-version}] [@b{--version}] [@b{-W}] [@b{--warn}]
a0b7da79
MM
244 [@b{--fatal-warnings}] [@b{-w}] [@b{-x}] [@b{-Z}] [@b{@@@var{FILE}}]
245 [@b{--target-help}] [@var{target-options}]
246 [@b{--}|@var{files} @dots{}]
a4fb0134
SC
247@c
248@c Target dependent options are listed below. Keep the list sorted.
01642c12 249@c Add an empty line for separation.
625e1353
RH
250@ifset ALPHA
251
252@emph{Target Alpha options:}
253 [@b{-m@var{cpu}}]
254 [@b{-mdebug} | @b{-no-mdebug}]
198f1251 255 [@b{-replace} | @b{-noreplace}]
625e1353
RH
256 [@b{-relax}] [@b{-g}] [@b{-G@var{size}}]
257 [@b{-F}] [@b{-32addr}]
258@end ifset
252b5132 259@ifset ARC
a4fb0134
SC
260
261@emph{Target ARC options:}
262 [@b{-marc[5|6|7|8]}]
263 [@b{-EB}|@b{-EL}]
252b5132
RH
264@end ifset
265@ifset ARM
a4fb0134
SC
266
267@emph{Target ARM options:}
03b1477f 268@c Don't document the deprecated options
92081f48
HPN
269 [@b{-mcpu}=@var{processor}[+@var{extension}@dots{}]]
270 [@b{-march}=@var{architecture}[+@var{extension}@dots{}]]
33a392fb
PB
271 [@b{-mfpu}=@var{floating-point-format}]
272 [@b{-mfloat-abi}=@var{abi}]
d507cf36 273 [@b{-meabi}=@var{ver}]
03b1477f 274 [@b{-mthumb}]
a4fb0134
SC
275 [@b{-EB}|@b{-EL}]
276 [@b{-mapcs-32}|@b{-mapcs-26}|@b{-mapcs-float}|
277 @b{-mapcs-reentrant}]
7f266840 278 [@b{-mthumb-interwork}] [@b{-k}]
252b5132 279@end ifset
9982501a
JZ
280@ifset Blackfin
281
282@emph{Target Blackfin options:}
283 [@b{-mcpu}=@var{processor}[-@var{sirevision}]]
284 [@b{-mfdpic}]
285 [@b{-mno-fdpic}]
286 [@b{-mnopic}]
287@end ifset
328eb32e
HPN
288@ifset CRIS
289
290@emph{Target CRIS options:}
291 [@b{--underscore} | @b{--no-underscore}]
292 [@b{--pic}] [@b{-N}]
293 [@b{--emulation=criself} | @b{--emulation=crisaout}]
ae57792d 294 [@b{--march=v0_v10} | @b{--march=v10} | @b{--march=v32} | @b{--march=common_v10_v32}]
328eb32e
HPN
295@c Deprecated -- deliberately not documented.
296@c [@b{-h}] [@b{-H}]
297@end ifset
252b5132 298@ifset D10V
a4fb0134
SC
299
300@emph{Target D10V options:}
301 [@b{-O}]
252b5132
RH
302@end ifset
303@ifset D30V
a4fb0134
SC
304
305@emph{Target D30V options:}
306 [@b{-O}|@b{-n}|@b{-N}]
252b5132
RH
307@end ifset
308@ifset H8
6fd4f6cc
DD
309
310@emph{Target H8/300 options:}
311 [-h-tick-hex]
252b5132
RH
312@end ifset
313@ifset HPPA
314@c HPPA has no machine-dependent assembler options (yet).
315@end ifset
a4fb0134
SC
316@ifset I80386
317
318@emph{Target i386 options:}
12b55ccc 319 [@b{--32}|@b{--64}] [@b{-n}]
1ef52f49 320 [@b{-march}=@var{CPU}[+@var{EXTENSION}@dots{}]] [@b{-mtune}=@var{CPU}]
252b5132
RH
321@end ifset
322@ifset I960
a4fb0134
SC
323
324@emph{Target i960 options:}
252b5132 325@c see md_parse_option in tc-i960.c
a4fb0134
SC
326 [@b{-ACA}|@b{-ACA_A}|@b{-ACB}|@b{-ACC}|@b{-AKA}|@b{-AKB}|
327 @b{-AKC}|@b{-AMC}]
328 [@b{-b}] [@b{-no-relax}]
252b5132 329@end ifset
587fe2b3 330@ifset IA64
a4fb0134 331
9e32ca89
NC
332@emph{Target IA-64 options:}
333 [@b{-mconstant-gp}|@b{-mauto-pic}]
334 [@b{-milp32}|@b{-milp64}|@b{-mlp64}|@b{-mp64}]
335 [@b{-mle}|@b{mbe}]
8c2fda1d 336 [@b{-mtune=itanium1}|@b{-mtune=itanium2}]
970d6792 337 [@b{-munwind-check=warning}|@b{-munwind-check=error}]
91d777ee 338 [@b{-mhint.b=ok}|@b{-mhint.b=warning}|@b{-mhint.b=error}]
9e32ca89
NC
339 [@b{-x}|@b{-xexplicit}] [@b{-xauto}] [@b{-xdebug}]
340@end ifset
a40cbfa3
NC
341@ifset IP2K
342
343@emph{Target IP2K options:}
344 [@b{-mip2022}|@b{-mip2022ext}]
345@end ifset
49f58d10
JB
346@ifset M32C
347
348@emph{Target M32C options:}
c54b5932 349 [@b{-m32c}|@b{-m16c}] [-relax] [-h-tick-hex]
49f58d10 350@end ifset
587fe2b3 351@ifset M32R
9e32ca89 352
a4fb0134
SC
353@emph{Target M32R options:}
354 [@b{--m32rx}|@b{--[no-]warn-explicit-parallel-conflicts}|
587fe2b3 355 @b{--W[n]p}]
ec694b89 356@end ifset
252b5132 357@ifset M680X0
a4fb0134
SC
358
359@emph{Target M680X0 options:}
360 [@b{-l}] [@b{-m68000}|@b{-m68010}|@b{-m68020}|@dots{}]
252b5132 361@end ifset
60bcf0fa 362@ifset M68HC11
a4fb0134
SC
363
364@emph{Target M68HC11 options:}
d01030e6 365 [@b{-m68hc11}|@b{-m68hc12}|@b{-m68hcs12}]
2f904664
SC
366 [@b{-mshort}|@b{-mlong}]
367 [@b{-mshort-double}|@b{-mlong-double}]
1370e33d 368 [@b{--force-long-branches}] [@b{--short-branches}]
a4fb0134
SC
369 [@b{--strict-direct-mode}] [@b{--print-insn-syntax}]
370 [@b{--print-opcodes}] [@b{--generate-example}]
371@end ifset
372@ifset MCORE
373
374@emph{Target MCORE options:}
375 [@b{-jsri2bsr}] [@b{-sifilter}] [@b{-relax}]
376 [@b{-mcpu=[210|340]}]
60bcf0fa 377@end ifset
7ba29e2a
NC
378@ifset MICROBLAZE
379@emph{Target MICROBLAZE options:}
380@c MicroBlaze has no machine-dependent assembler options.
381@end ifset
252b5132 382@ifset MIPS
a4fb0134
SC
383
384@emph{Target MIPS options:}
78849248 385 [@b{-nocpp}] [@b{-EL}] [@b{-EB}] [@b{-O}[@var{optimization level}]]
437ee9d5 386 [@b{-g}[@var{debug level}]] [@b{-G} @var{num}] [@b{-KPIC}] [@b{-call_shared}]
0c000745 387 [@b{-non_shared}] [@b{-xgot} [@b{-mvxworks-pic}]
437ee9d5
TS
388 [@b{-mabi}=@var{ABI}] [@b{-32}] [@b{-n32}] [@b{-64}] [@b{-mfp32}] [@b{-mgp32}]
389 [@b{-march}=@var{CPU}] [@b{-mtune}=@var{CPU}] [@b{-mips1}] [@b{-mips2}]
af7ee8bf 390 [@b{-mips3}] [@b{-mips4}] [@b{-mips5}] [@b{-mips32}] [@b{-mips32r2}]
5f74bc13 391 [@b{-mips64}] [@b{-mips64r2}]
437ee9d5
TS
392 [@b{-construct-floats}] [@b{-no-construct-floats}]
393 [@b{-trap}] [@b{-no-break}] [@b{-break}] [@b{-no-trap}]
437ee9d5 394 [@b{-mips16}] [@b{-no-mips16}]
e16bfa71 395 [@b{-msmartmips}] [@b{-mno-smartmips}]
1f25f5d3 396 [@b{-mips3d}] [@b{-no-mips3d}]
deec1734 397 [@b{-mdmx}] [@b{-no-mdmx}]
2ef2b9ae 398 [@b{-mdsp}] [@b{-mno-dsp}]
8b082fb1 399 [@b{-mdspr2}] [@b{-mno-dspr2}]
ef2e4d86 400 [@b{-mmt}] [@b{-mno-mt}]
2babba43
MR
401 [@b{-mfix7000}] [@b{-mno-fix7000}]
402 [@b{-mfix-vr4120}] [@b{-mno-fix-vr4120}]
403 [@b{-mfix-vr4130}] [@b{-mno-fix-vr4130}]
ecb4347a 404 [@b{-mdebug}] [@b{-no-mdebug}]
dcd410fe 405 [@b{-mpdr}] [@b{-mno-pdr}]
3c3bdf30
NC
406@end ifset
407@ifset MMIX
a4fb0134
SC
408
409@emph{Target MMIX options:}
410 [@b{--fixed-special-register-names}] [@b{--globalize-symbols}]
411 [@b{--gnu-syntax}] [@b{--relax}] [@b{--no-predefined-symbols}]
412 [@b{--no-expand}] [@b{--no-merge-gregs}] [@b{-x}]
973eb340 413 [@b{--linker-allocated-gregs}]
a4fb0134
SC
414@end ifset
415@ifset PDP11
416
417@emph{Target PDP11 options:}
418 [@b{-mpic}|@b{-mno-pic}] [@b{-mall}] [@b{-mno-extensions}]
419 [@b{-m}@var{extension}|@b{-mno-}@var{extension}]
01642c12 420 [@b{-m}@var{cpu}] [@b{-m}@var{machine}]
a4fb0134
SC
421@end ifset
422@ifset PJ
423
424@emph{Target picoJava options:}
425 [@b{-mb}|@b{-me}]
426@end ifset
427@ifset PPC
428
429@emph{Target PowerPC options:}
430 [@b{-mpwrx}|@b{-mpwr2}|@b{-mpwr}|@b{-m601}|@b{-mppc}|@b{-mppc32}|@b{-m603}|@b{-m604}|
2f3bb96a 431 @b{-m403}|@b{-m405}|@b{-mppc64}|@b{-m620}|@b{-mppc64bridge}|@b{-mbooke}]
9b4e5766 432 [@b{-mcom}|@b{-many}|@b{-maltivec}|@b{-mvsx}] [@b{-memb}]
a4fb0134
SC
433 [@b{-mregnames}|@b{-mno-regnames}]
434 [@b{-mrelocatable}|@b{-mrelocatable-lib}]
435 [@b{-mlittle}|@b{-mlittle-endian}|@b{-mbig}|@b{-mbig-endian}]
436 [@b{-msolaris}|@b{-mno-solaris}]
437@end ifset
c7927a3c
NC
438@ifset RX
439
440@emph{Target RX options:}
441 [@b{-mlittle-endian}|@b{-mbig-endian}]
442 [@b{-m32bit-ints}|@b{-m16bit-ints}]
443 [@b{-m32bit-doubles}|@b{-m64bit-doubles}]
444@end ifset
11c19e16
MS
445@ifset S390
446
447@emph{Target s390 options:}
448 [@b{-m31}|@b{-m64}] [@b{-mesa}|@b{-mzarch}] [@b{-march}=@var{CPU}]
449 [@b{-mregnames}|@b{-mno-regnames}]
450 [@b{-mwarn-areg-zero}]
451@end ifset
c3b7224a
NC
452@ifset SCORE
453
454@emph{Target SCORE options:}
455 [@b{-EB}][@b{-EL}][@b{-FIXDD}][@b{-NWARN}]
456 [@b{-SCORE5}][@b{-SCORE5U}][@b{-SCORE7}][@b{-SCORE3}]
457 [@b{-march=score7}][@b{-march=score3}]
458 [@b{-USE_R1}][@b{-KPIC}][@b{-O0}][@b{-G} @var{num}][@b{-V}]
459@end ifset
a4fb0134
SC
460@ifset SPARC
461
462@emph{Target SPARC options:}
463@c The order here is important. See c-sparc.texi.
464 [@b{-Av6}|@b{-Av7}|@b{-Av8}|@b{-Asparclet}|@b{-Asparclite}
465 @b{-Av8plus}|@b{-Av8plusa}|@b{-Av9}|@b{-Av9a}]
466 [@b{-xarch=v8plus}|@b{-xarch=v8plusa}] [@b{-bump}]
467 [@b{-32}|@b{-64}]
468@end ifset
469@ifset TIC54X
470
471@emph{Target TIC54X options:}
01642c12 472 [@b{-mcpu=54[123589]}|@b{-mcpu=54[56]lp}] [@b{-mfar-mode}|@b{-mf}]
a4fb0134
SC
473 [@b{-merrors-to-file} @var{<filename>}|@b{-me} @var{<filename>}]
474@end ifset
3c9b82ba 475
40b36596
JM
476@ifset TIC6X
477
478@emph{Target TIC6X options:}
479 [@b{-march=@var{arch}}] [@b{-matomic}|@b{-mno-atomic}]
480 [@b{-mbig-endian}|@b{-mlittle-endian}]
481@end ifset
482
3c9b82ba
NC
483@ifset Z80
484
485@emph{Target Z80 options:}
486 [@b{-z80}] [@b{-r800}]
487 [@b{ -ignore-undocumented-instructions}] [@b{-Wnud}]
488 [@b{ -ignore-unportable-instructions}] [@b{-Wnup}]
489 [@b{ -warn-undocumented-instructions}] [@b{-Wud}]
490 [@b{ -warn-unportable-instructions}] [@b{-Wup}]
491 [@b{ -forbid-undocumented-instructions}] [@b{-Fud}]
492 [@b{ -forbid-unportable-instructions}] [@b{-Fup}]
493@end ifset
494
a4fb0134
SC
495@ifset Z8000
496@c Z8000 has no machine-dependent assembler options
252b5132 497@end ifset
e0001a05
NC
498@ifset XTENSA
499
500@emph{Target Xtensa options:}
43cd72b9 501 [@b{--[no-]text-section-literals}] [@b{--[no-]absolute-literals}]
e0001a05 502 [@b{--[no-]target-align}] [@b{--[no-]longcalls}]
43cd72b9 503 [@b{--[no-]transform}]
9456465c 504 [@b{--rename-section} @var{oldname}=@var{newname}]
e0001a05 505@end ifset
0285c67d 506@c man end
252b5132
RH
507@end smallexample
508
0285c67d
NC
509@c man begin OPTIONS
510
a4fb0134 511@table @gcctabopt
38fc1cb1 512@include at-file.texi
a0b7da79 513
83f10cb2 514@item -a[cdghlmns]
252b5132
RH
515Turn on listings, in any of a variety of ways:
516
a4fb0134 517@table @gcctabopt
252b5132
RH
518@item -ac
519omit false conditionals
520
521@item -ad
522omit debugging directives
523
83f10cb2
NC
524@item -ag
525include general information, like @value{AS} version and options passed
526
252b5132
RH
527@item -ah
528include high-level source
529
530@item -al
531include assembly
532
533@item -am
534include macro expansions
535
536@item -an
537omit forms processing
538
539@item -as
540include symbols
541
542@item =file
543set the name of the listing file
544@end table
545
546You may combine these options; for example, use @samp{-aln} for assembly
547listing without forms processing. The @samp{=file} option, if used, must be
548the last one. By itself, @samp{-a} defaults to @samp{-ahls}.
549
caa32fe5 550@item --alternate
96e9638b
BW
551Begin in alternate macro mode.
552@ifclear man
553@xref{Altmacro,,@code{.altmacro}}.
554@end ifclear
caa32fe5 555
955974c6
CC
556@item --compress-debug-sections
557Compress DWARF debug sections using zlib. The debug sections are renamed
558to begin with @samp{.zdebug}, and the resulting object file may not be
559compatible with older linkers and object file utilities.
560
561@item --nocompress-debug-sections
562Do not compress DWARF debug sections. This is the default.
563
252b5132
RH
564@item -D
565Ignored. This option is accepted for script compatibility with calls to
566other assemblers.
567
3d6b762c
JM
568@item --debug-prefix-map @var{old}=@var{new}
569When assembling files in directory @file{@var{old}}, record debugging
570information describing them as in @file{@var{new}} instead.
571
252b5132
RH
572@item --defsym @var{sym}=@var{value}
573Define the symbol @var{sym} to be @var{value} before assembling the input file.
574@var{value} must be an integer constant. As in C, a leading @samp{0x}
bf083c64
NC
575indicates a hexadecimal value, and a leading @samp{0} indicates an octal
576value. The value of the symbol can be overridden inside a source file via the
577use of a @code{.set} pseudo-op.
252b5132
RH
578
579@item -f
580``fast''---skip whitespace and comment preprocessing (assume source is
581compiler output).
582
329e276d
NC
583@item -g
584@itemx --gen-debug
585Generate debugging information for each assembler source line using whichever
586debug format is preferred by the target. This currently means either STABS,
587ECOFF or DWARF2.
588
252b5132
RH
589@item --gstabs
590Generate stabs debugging information for each assembler line. This
591may help debugging assembler code, if the debugger can handle it.
592
05da4302
NC
593@item --gstabs+
594Generate stabs debugging information for each assembler line, with GNU
595extensions that probably only gdb can handle, and that could make other
596debuggers crash or refuse to read your program. This
597may help debugging assembler code. Currently the only GNU extension is
598the location of the current working directory at assembling time.
599
329e276d 600@item --gdwarf-2
cdf82bcf 601Generate DWARF2 debugging information for each assembler line. This
c1253627 602may help debugging assembler code, if the debugger can handle it. Note---this
85a39694 603option is only supported by some targets, not all of them.
cdf82bcf 604
252b5132
RH
605@item --help
606Print a summary of the command line options and exit.
607
ea20a7da
CC
608@item --target-help
609Print a summary of all target specific options and exit.
610
252b5132
RH
611@item -I @var{dir}
612Add directory @var{dir} to the search list for @code{.include} directives.
613
614@item -J
615Don't warn about signed overflow.
616
617@item -K
618@ifclear DIFF-TBL-KLUGE
619This option is accepted but has no effect on the @value{TARGET} family.
620@end ifclear
621@ifset DIFF-TBL-KLUGE
622Issue warnings when difference tables altered for long displacements.
623@end ifset
624
625@item -L
626@itemx --keep-locals
ba83aca1
BW
627Keep (in the symbol table) local symbols. These symbols start with
628system-specific local label prefixes, typically @samp{.L} for ELF systems
629or @samp{L} for traditional a.out systems.
630@ifclear man
631@xref{Symbol Names}.
632@end ifclear
252b5132 633
c3a27914
NC
634@item --listing-lhs-width=@var{number}
635Set the maximum width, in words, of the output data column for an assembler
636listing to @var{number}.
637
638@item --listing-lhs-width2=@var{number}
639Set the maximum width, in words, of the output data column for continuation
640lines in an assembler listing to @var{number}.
641
642@item --listing-rhs-width=@var{number}
643Set the maximum width of an input source line, as displayed in a listing, to
644@var{number} bytes.
645
646@item --listing-cont-lines=@var{number}
647Set the maximum number of lines printed in a listing for a single line of input
648to @var{number} + 1.
649
252b5132 650@item -o @var{objfile}
a4fb0134 651Name the object-file output from @command{@value{AS}} @var{objfile}.
252b5132
RH
652
653@item -R
654Fold the data section into the text section.
655
4bdd3565
NC
656@kindex --hash-size=@var{number}
657Set the default size of GAS's hash tables to a prime number close to
658@var{number}. Increasing this value can reduce the length of time it takes the
659assembler to perform its tasks, at the expense of increasing the assembler's
660memory requirements. Similarly reducing this value can reduce the memory
661requirements at the expense of speed.
662
663@item --reduce-memory-overheads
664This option reduces GAS's memory requirements, at the expense of making the
665assembly processes slower. Currently this switch is a synonym for
666@samp{--hash-size=4051}, but in the future it may have other effects as well.
667
252b5132
RH
668@item --statistics
669Print the maximum space (in bytes) and total time (in seconds) used by
670assembly.
671
672@item --strip-local-absolute
673Remove local absolute symbols from the outgoing symbol table.
674
675@item -v
676@itemx -version
a4fb0134 677Print the @command{as} version.
252b5132
RH
678
679@item --version
a4fb0134 680Print the @command{as} version and exit.
252b5132
RH
681
682@item -W
2bdd6cf5 683@itemx --no-warn
252b5132
RH
684Suppress warning messages.
685
2bdd6cf5
GK
686@item --fatal-warnings
687Treat warnings as errors.
688
689@item --warn
690Don't suppress warning messages or treat them as errors.
691
252b5132
RH
692@item -w
693Ignored.
694
695@item -x
696Ignored.
697
698@item -Z
699Generate an object file even after errors.
700
701@item -- | @var{files} @dots{}
702Standard input, or source files to assemble.
703
704@end table
705
706@ifset ARC
707The following options are available when @value{AS} is configured for
708an ARC processor.
709
a4fb0134 710@table @gcctabopt
0d2bcfaf
NC
711@item -marc[5|6|7|8]
712This option selects the core processor variant.
713@item -EB | -EL
714Select either big-endian (-EB) or little-endian (-EL) output.
252b5132
RH
715@end table
716@end ifset
717
718@ifset ARM
719The following options are available when @value{AS} is configured for the ARM
720processor family.
721
a4fb0134 722@table @gcctabopt
92081f48 723@item -mcpu=@var{processor}[+@var{extension}@dots{}]
cdf82bcf 724Specify which ARM processor variant is the target.
92081f48 725@item -march=@var{architecture}[+@var{extension}@dots{}]
cdf82bcf 726Specify which ARM architecture variant is used by the target.
03b1477f 727@item -mfpu=@var{floating-point-format}
a349d9dd 728Select which Floating Point architecture is the target.
33a392fb
PB
729@item -mfloat-abi=@var{abi}
730Select which floating point ABI is in use.
03b1477f
RE
731@item -mthumb
732Enable Thumb only instruction decoding.
7f266840 733@item -mapcs-32 | -mapcs-26 | -mapcs-float | -mapcs-reentrant
252b5132
RH
734Select which procedure calling convention is in use.
735@item -EB | -EL
736Select either big-endian (-EB) or little-endian (-EL) output.
cdf82bcf
NC
737@item -mthumb-interwork
738Specify that the code has been generated with interworking between Thumb and
739ARM code in mind.
740@item -k
741Specify that PIC code has been generated.
252b5132
RH
742@end table
743@end ifset
744
9982501a
JZ
745@ifset Blackfin
746The following options are available when @value{AS} is configured for
747the Blackfin processor family.
748
749@table @gcctabopt
750@item -mcpu=@var{processor}@r{[}-@var{sirevision}@r{]}
751This option specifies the target processor. The optional @var{sirevision}
752is not used in assembler.
753@item -mfdpic
754Assemble for the FDPIC ABI.
1f9bb1ca
AS
755@item -mno-fdpic
756@itemx -mnopic
9982501a
JZ
757Disable -mfdpic.
758@end table
759@end ifset
760
328eb32e
HPN
761@ifset CRIS
762See the info pages for documentation of the CRIS-specific options.
763@end ifset
764
252b5132
RH
765@ifset D10V
766The following options are available when @value{AS} is configured for
767a D10V processor.
a4fb0134 768@table @gcctabopt
252b5132
RH
769@cindex D10V optimization
770@cindex optimization, D10V
771@item -O
772Optimize output by parallelizing instructions.
773@end table
774@end ifset
775
776@ifset D30V
777The following options are available when @value{AS} is configured for a D30V
778processor.
a4fb0134 779@table @gcctabopt
252b5132
RH
780@cindex D30V optimization
781@cindex optimization, D30V
782@item -O
783Optimize output by parallelizing instructions.
784
785@cindex D30V nops
786@item -n
787Warn when nops are generated.
788
789@cindex D30V nops after 32-bit multiply
790@item -N
791Warn when a nop after a 32-bit multiply instruction is generated.
792@end table
793@end ifset
794
795@ifset I960
796The following options are available when @value{AS} is configured for the
797Intel 80960 processor.
798
a4fb0134 799@table @gcctabopt
252b5132
RH
800@item -ACA | -ACA_A | -ACB | -ACC | -AKA | -AKB | -AKC | -AMC
801Specify which variant of the 960 architecture is the target.
802
803@item -b
804Add code to collect statistics about branches taken.
805
806@item -no-relax
807Do not alter compare-and-branch instructions for long displacements;
808error if necessary.
809
810@end table
811@end ifset
812
a40cbfa3
NC
813@ifset IP2K
814The following options are available when @value{AS} is configured for the
ec88d317 815Ubicom IP2K series.
a40cbfa3
NC
816
817@table @gcctabopt
818
819@item -mip2022ext
820Specifies that the extended IP2022 instructions are allowed.
821
822@item -mip2022
8dfa0188 823Restores the default behaviour, which restricts the permitted instructions to
a40cbfa3
NC
824just the basic IP2022 ones.
825
826@end table
827@end ifset
828
49f58d10
JB
829@ifset M32C
830The following options are available when @value{AS} is configured for the
831Renesas M32C and M16C processors.
832
833@table @gcctabopt
834
835@item -m32c
836Assemble M32C instructions.
837
838@item -m16c
839Assemble M16C instructions (the default).
840
c54b5932
DD
841@item -relax
842Enable support for link-time relaxations.
843
844@item -h-tick-hex
845Support H'00 style hex constants in addition to 0x00 style.
846
49f58d10
JB
847@end table
848@end ifset
849
ec694b89
NC
850@ifset M32R
851The following options are available when @value{AS} is configured for the
26597c86 852Renesas M32R (formerly Mitsubishi M32R) series.
ec694b89 853
a4fb0134 854@table @gcctabopt
ec694b89
NC
855
856@item --m32rx
857Specify which processor in the M32R family is the target. The default
858is normally the M32R, but this option changes it to the M32RX.
859
860@item --warn-explicit-parallel-conflicts or --Wp
861Produce warning messages when questionable parallel constructs are
01642c12 862encountered.
ec694b89
NC
863
864@item --no-warn-explicit-parallel-conflicts or --Wnp
01642c12
RM
865Do not produce warning messages when questionable parallel constructs are
866encountered.
ec694b89
NC
867
868@end table
869@end ifset
252b5132
RH
870
871@ifset M680X0
872The following options are available when @value{AS} is configured for the
873Motorola 68000 series.
874
a4fb0134 875@table @gcctabopt
252b5132
RH
876
877@item -l
878Shorten references to undefined symbols, to one word instead of two.
879
0285c67d
NC
880@item -m68000 | -m68008 | -m68010 | -m68020 | -m68030
881@itemx | -m68040 | -m68060 | -m68302 | -m68331 | -m68332
882@itemx | -m68333 | -m68340 | -mcpu32 | -m5200
252b5132
RH
883Specify what processor in the 68000 family is the target. The default
884is normally the 68020, but this can be changed at configuration time.
885
886@item -m68881 | -m68882 | -mno-68881 | -mno-68882
887The target machine does (or does not) have a floating-point coprocessor.
888The default is to assume a coprocessor for 68020, 68030, and cpu32. Although
889the basic 68000 is not compatible with the 68881, a combination of the
890two can be specified, since it's possible to do emulation of the
891coprocessor instructions with the main processor.
892
893@item -m68851 | -mno-68851
894The target machine does (or does not) have a memory-management
895unit coprocessor. The default is to assume an MMU for 68020 and up.
896
897@end table
898@end ifset
899
e135f41b
NC
900@ifset PDP11
901
902For details about the PDP-11 machine dependent features options,
903see @ref{PDP-11-Options}.
904
a4fb0134 905@table @gcctabopt
e135f41b
NC
906@item -mpic | -mno-pic
907Generate position-independent (or position-dependent) code. The
a4fb0134 908default is @option{-mpic}.
e135f41b
NC
909
910@item -mall
911@itemx -mall-extensions
912Enable all instruction set extensions. This is the default.
913
914@item -mno-extensions
915Disable all instruction set extensions.
916
917@item -m@var{extension} | -mno-@var{extension}
918Enable (or disable) a particular instruction set extension.
919
920@item -m@var{cpu}
921Enable the instruction set extensions supported by a particular CPU, and
922disable all other extensions.
923
924@item -m@var{machine}
925Enable the instruction set extensions supported by a particular machine
926model, and disable all other extensions.
927@end table
928
929@end ifset
930
041dd5a9
ILT
931@ifset PJ
932The following options are available when @value{AS} is configured for
933a picoJava processor.
934
a4fb0134 935@table @gcctabopt
041dd5a9
ILT
936
937@cindex PJ endianness
938@cindex endianness, PJ
939@cindex big endian output, PJ
940@item -mb
941Generate ``big endian'' format output.
942
943@cindex little endian output, PJ
944@item -ml
945Generate ``little endian'' format output.
946
947@end table
948@end ifset
949
60bcf0fa
NC
950@ifset M68HC11
951The following options are available when @value{AS} is configured for the
952Motorola 68HC11 or 68HC12 series.
953
a4fb0134 954@table @gcctabopt
60bcf0fa 955
d01030e6 956@item -m68hc11 | -m68hc12 | -m68hcs12
60bcf0fa
NC
957Specify what processor is the target. The default is
958defined by the configuration option when building the assembler.
959
2f904664
SC
960@item -mshort
961Specify to use the 16-bit integer ABI.
962
963@item -mlong
01642c12 964Specify to use the 32-bit integer ABI.
2f904664
SC
965
966@item -mshort-double
01642c12 967Specify to use the 32-bit double ABI.
2f904664
SC
968
969@item -mlong-double
01642c12 970Specify to use the 64-bit double ABI.
2f904664 971
1370e33d 972@item --force-long-branches
60bcf0fa
NC
973Relative branches are turned into absolute ones. This concerns
974conditional branches, unconditional branches and branches to a
975sub routine.
976
1370e33d
NC
977@item -S | --short-branches
978Do not turn relative branches into absolute ones
60bcf0fa
NC
979when the offset is out of range.
980
981@item --strict-direct-mode
982Do not turn the direct addressing mode into extended addressing mode
983when the instruction does not support direct addressing mode.
984
985@item --print-insn-syntax
986Print the syntax of instruction in case of error.
987
988@item --print-opcodes
989print the list of instructions with syntax and then exit.
990
991@item --generate-example
992print an example of instruction for each possible instruction and then exit.
a4fb0134 993This option is only useful for testing @command{@value{AS}}.
60bcf0fa
NC
994
995@end table
996@end ifset
997
252b5132 998@ifset SPARC
a4fb0134 999The following options are available when @command{@value{AS}} is configured
252b5132
RH
1000for the SPARC architecture:
1001
a4fb0134 1002@table @gcctabopt
252b5132
RH
1003@item -Av6 | -Av7 | -Av8 | -Asparclet | -Asparclite
1004@itemx -Av8plus | -Av8plusa | -Av9 | -Av9a
1005Explicitly select a variant of the SPARC architecture.
1006
1007@samp{-Av8plus} and @samp{-Av8plusa} select a 32 bit environment.
1008@samp{-Av9} and @samp{-Av9a} select a 64 bit environment.
1009
1010@samp{-Av8plusa} and @samp{-Av9a} enable the SPARC V9 instruction set with
1011UltraSPARC extensions.
1012
1013@item -xarch=v8plus | -xarch=v8plusa
1014For compatibility with the Solaris v9 assembler. These options are
1015equivalent to -Av8plus and -Av8plusa, respectively.
1016
1017@item -bump
1018Warn when the assembler switches to another architecture.
1019@end table
1020@end ifset
1021
39bec121
TW
1022@ifset TIC54X
1023The following options are available when @value{AS} is configured for the 'c54x
01642c12 1024architecture.
39bec121 1025
a4fb0134 1026@table @gcctabopt
39bec121
TW
1027@item -mfar-mode
1028Enable extended addressing mode. All addresses and relocations will assume
1029extended addressing (usually 23 bits).
1030@item -mcpu=@var{CPU_VERSION}
1031Sets the CPU version being compiled for.
1032@item -merrors-to-file @var{FILENAME}
1033Redirect error output to a file, for broken systems which don't support such
1034behaviour in the shell.
1035@end table
1036@end ifset
1037
252b5132
RH
1038@ifset MIPS
1039The following options are available when @value{AS} is configured for
437ee9d5 1040a @sc{mips} processor.
252b5132 1041
a4fb0134 1042@table @gcctabopt
252b5132
RH
1043@item -G @var{num}
1044This option sets the largest size of an object that can be referenced
1045implicitly with the @code{gp} register. It is only accepted for targets that
1046use ECOFF format, such as a DECstation running Ultrix. The default value is 8.
1047
1048@cindex MIPS endianness
1049@cindex endianness, MIPS
1050@cindex big endian output, MIPS
1051@item -EB
1052Generate ``big endian'' format output.
1053
1054@cindex little endian output, MIPS
1055@item -EL
1056Generate ``little endian'' format output.
1057
1058@cindex MIPS ISA
1059@item -mips1
1060@itemx -mips2
1061@itemx -mips3
e7af610e 1062@itemx -mips4
437ee9d5 1063@itemx -mips5
e7af610e 1064@itemx -mips32
af7ee8bf 1065@itemx -mips32r2
4058e45f 1066@itemx -mips64
5f74bc13 1067@itemx -mips64r2
437ee9d5
TS
1068Generate code for a particular @sc{mips} Instruction Set Architecture level.
1069@samp{-mips1} is an alias for @samp{-march=r3000}, @samp{-mips2} is an
1070alias for @samp{-march=r6000}, @samp{-mips3} is an alias for
1071@samp{-march=r4000} and @samp{-mips4} is an alias for @samp{-march=r8000}.
5f74bc13
CD
1072@samp{-mips5}, @samp{-mips32}, @samp{-mips32r2}, @samp{-mips64}, and
1073@samp{-mips64r2}
af7ee8bf 1074correspond to generic
5f74bc13
CD
1075@samp{MIPS V}, @samp{MIPS32}, @samp{MIPS32 Release 2}, @samp{MIPS64},
1076and @samp{MIPS64 Release 2}
1077ISA processors, respectively.
437ee9d5
TS
1078
1079@item -march=@var{CPU}
1080Generate code for a particular @sc{mips} cpu.
1081
1082@item -mtune=@var{cpu}
1083Schedule and tune for a particular @sc{mips} cpu.
1084
1085@item -mfix7000
1086@itemx -mno-fix7000
1087Cause nops to be inserted if the read of the destination register
1088of an mfhi or mflo instruction occurs in the following two instructions.
1089
ecb4347a
DJ
1090@item -mdebug
1091@itemx -no-mdebug
1092Cause stabs-style debugging output to go into an ECOFF-style .mdebug
1093section instead of the standard ELF .stabs sections.
1094
dcd410fe
RO
1095@item -mpdr
1096@itemx -mno-pdr
1097Control generation of @code{.pdr} sections.
1098
437ee9d5
TS
1099@item -mgp32
1100@itemx -mfp32
1101The register sizes are normally inferred from the ISA and ABI, but these
1102flags force a certain group of registers to be treated as 32 bits wide at
1103all times. @samp{-mgp32} controls the size of general-purpose registers
1104and @samp{-mfp32} controls the size of floating-point registers.
1105
1106@item -mips16
1107@itemx -no-mips16
1108Generate code for the MIPS 16 processor. This is equivalent to putting
1109@code{.set mips16} at the start of the assembly file. @samp{-no-mips16}
1110turns off this option.
252b5132 1111
e16bfa71
TS
1112@item -msmartmips
1113@itemx -mno-smartmips
1114Enables the SmartMIPS extension to the MIPS32 instruction set. This is
1115equivalent to putting @code{.set smartmips} at the start of the assembly file.
1116@samp{-mno-smartmips} turns off this option.
1117
1f25f5d3
CD
1118@item -mips3d
1119@itemx -no-mips3d
1120Generate code for the MIPS-3D Application Specific Extension.
1121This tells the assembler to accept MIPS-3D instructions.
1122@samp{-no-mips3d} turns off this option.
1123
deec1734
CD
1124@item -mdmx
1125@itemx -no-mdmx
1126Generate code for the MDMX Application Specific Extension.
1127This tells the assembler to accept MDMX instructions.
1128@samp{-no-mdmx} turns off this option.
1129
2ef2b9ae
CF
1130@item -mdsp
1131@itemx -mno-dsp
8b082fb1
TS
1132Generate code for the DSP Release 1 Application Specific Extension.
1133This tells the assembler to accept DSP Release 1 instructions.
2ef2b9ae
CF
1134@samp{-mno-dsp} turns off this option.
1135
8b082fb1
TS
1136@item -mdspr2
1137@itemx -mno-dspr2
1138Generate code for the DSP Release 2 Application Specific Extension.
1139This option implies -mdsp.
1140This tells the assembler to accept DSP Release 2 instructions.
1141@samp{-mno-dspr2} turns off this option.
1142
ef2e4d86
CF
1143@item -mmt
1144@itemx -mno-mt
1145Generate code for the MT Application Specific Extension.
1146This tells the assembler to accept MT instructions.
1147@samp{-mno-mt} turns off this option.
1148
437ee9d5
TS
1149@item --construct-floats
1150@itemx --no-construct-floats
1151The @samp{--no-construct-floats} option disables the construction of
1152double width floating point constants by loading the two halves of the
1153value into the two single width floating point registers that make up
1154the double width register. By default @samp{--construct-floats} is
1155selected, allowing construction of these floating point constants.
252b5132
RH
1156
1157@cindex emulation
1158@item --emulation=@var{name}
a4fb0134 1159This option causes @command{@value{AS}} to emulate @command{@value{AS}} configured
252b5132
RH
1160for some other target, in all respects, including output format (choosing
1161between ELF and ECOFF only), handling of pseudo-opcodes which may generate
1162debugging information or store symbol table information, and default
1163endianness. The available configuration names are: @samp{mipsecoff},
1164@samp{mipself}, @samp{mipslecoff}, @samp{mipsbecoff}, @samp{mipslelf},
1165@samp{mipsbelf}. The first two do not alter the default endianness from that
1166of the primary target for which the assembler was configured; the others change
1167the default to little- or big-endian as indicated by the @samp{b} or @samp{l}
1168in the name. Using @samp{-EB} or @samp{-EL} will override the endianness
1169selection in any case.
1170
1171This option is currently supported only when the primary target
437ee9d5 1172@command{@value{AS}} is configured for is a @sc{mips} ELF or ECOFF target.
252b5132
RH
1173Furthermore, the primary target or others specified with
1174@samp{--enable-targets=@dots{}} at configuration time must include support for
1175the other format, if both are to be available. For example, the Irix 5
1176configuration includes support for both.
1177
1178Eventually, this option will support more configurations, with more
1179fine-grained control over the assembler's behavior, and will be supported for
1180more processors.
1181
1182@item -nocpp
a4fb0134 1183@command{@value{AS}} ignores this option. It is accepted for compatibility with
252b5132
RH
1184the native tools.
1185
252b5132
RH
1186@item --trap
1187@itemx --no-trap
1188@itemx --break
1189@itemx --no-break
1190Control how to deal with multiplication overflow and division by zero.
1191@samp{--trap} or @samp{--no-break} (which are synonyms) take a trap exception
1192(and only work for Instruction Set Architecture level 2 and higher);
1193@samp{--break} or @samp{--no-trap} (also synonyms, and the default) take a
1194break exception.
63486801
L
1195
1196@item -n
a4fb0134 1197When this option is used, @command{@value{AS}} will issue a warning every
63486801 1198time it generates a nop instruction from a macro.
252b5132
RH
1199@end table
1200@end ifset
1201
1202@ifset MCORE
1203The following options are available when @value{AS} is configured for
1204an MCore processor.
1205
a4fb0134 1206@table @gcctabopt
252b5132
RH
1207@item -jsri2bsr
1208@itemx -nojsri2bsr
1209Enable or disable the JSRI to BSR transformation. By default this is enabled.
1210The command line option @samp{-nojsri2bsr} can be used to disable it.
1211
1212@item -sifilter
1213@itemx -nosifilter
1214Enable or disable the silicon filter behaviour. By default this is disabled.
a349d9dd 1215The default can be overridden by the @samp{-sifilter} command line option.
252b5132
RH
1216
1217@item -relax
1218Alter jump instructions for long displacements.
1219
ec694b89
NC
1220@item -mcpu=[210|340]
1221Select the cpu type on the target hardware. This controls which instructions
1222can be assembled.
1223
1224@item -EB
1225Assemble for a big endian target.
1226
1227@item -EL
1228Assemble for a little endian target.
252b5132
RH
1229
1230@end table
1231@end ifset
1232
3c3bdf30
NC
1233@ifset MMIX
1234See the info pages for documentation of the MMIX-specific options.
1235@end ifset
1236
046d31c2
NC
1237@ifset RX
1238See the info pages for documentation of the RX-specific options.
1239@end ifset
1240
11c19e16
MS
1241@ifset S390
1242The following options are available when @value{AS} is configured for the s390
1243processor family.
1244
1245@table @gcctabopt
1246@item -m31
1247@itemx -m64
1248Select the word size, either 31/32 bits or 64 bits.
1249@item -mesa
1250@item -mzarch
1251Select the architecture mode, either the Enterprise System
1252Architecture (esa) or the z/Architecture mode (zarch).
1253@item -march=@var{processor}
1254Specify which s390 processor variant is the target, @samp{g6}, @samp{g6},
1255@samp{z900}, @samp{z990}, @samp{z9-109}, @samp{z9-ec}, or @samp{z10}.
1256@item -mregnames
1257@itemx -mno-regnames
1258Allow or disallow symbolic names for registers.
1259@item -mwarn-areg-zero
1260Warn whenever the operand for a base or index register has been specified
1261but evaluates to zero.
1262@end table
1263@end ifset
1264
40b36596
JM
1265@ifset TIC6X
1266The following options are available when @value{AS} is configured for a
1267TMS320C6000 processor.
1268
1269@table @gcctabopt
1270@item -march=@var{arch}
1271Enable (only) instructions from architecture @var{arch}. By default,
1272all instructions are permitted.
1273
1274The following values of @var{arch} are accepted: @code{c62x},
1275@code{c64x}, @code{c64x+}, @code{c67x}, @code{c67x+}, @code{c674x}.
1276
1277@item -matomic
1278@itemx -mno-atomic
1279Enable or disable the optional C64x+ atomic operation instructions.
1280By default, they are enabled if no @option{-march} option is given, or
1281if an architecture is specified with @option{-march} that implies
1282these instructions are present (currently, there are no such
1283architectures); they are disabled if an architecture is specified with
1284@option{-march} on which the instructions are optional or not
1285present. This option overrides such a default from the architecture,
1286independent of the order in which the @option{-march} or
1287@option{-matomic} or @option{-mno-atomic} options are passed.
1288
1289@item -mbig-endian
1290@itemx -mlittle-endian
1291Generate code for the specified endianness. The default is
1292little-endian.
1293@end table
1294
1295@end ifset
1296
e0001a05
NC
1297@ifset XTENSA
1298The following options are available when @value{AS} is configured for
1299an Xtensa processor.
1300
1301@table @gcctabopt
e0001a05
NC
1302@item --text-section-literals | --no-text-section-literals
1303With @option{--text-@-section-@-literals}, literal pools are interspersed
1304in the text section. The default is
1305@option{--no-@-text-@-section-@-literals}, which places literals in a
43cd72b9
BW
1306separate section in the output file. These options only affect literals
1307referenced via PC-relative @code{L32R} instructions; literals for
1308absolute mode @code{L32R} instructions are handled separately.
1309
1310@item --absolute-literals | --no-absolute-literals
1311Indicate to the assembler whether @code{L32R} instructions use absolute
1312or PC-relative addressing. The default is to assume absolute addressing
1313if the Xtensa processor includes the absolute @code{L32R} addressing
1314option. Otherwise, only the PC-relative @code{L32R} mode can be used.
e0001a05
NC
1315
1316@item --target-align | --no-target-align
1317Enable or disable automatic alignment to reduce branch penalties at the
1318expense of some code density. The default is @option{--target-@-align}.
1319
1320@item --longcalls | --no-longcalls
1321Enable or disable transformation of call instructions to allow calls
1322across a greater range of addresses. The default is
1323@option{--no-@-longcalls}.
43cd72b9
BW
1324
1325@item --transform | --no-transform
1326Enable or disable all assembler transformations of Xtensa instructions.
1327The default is @option{--transform};
1328@option{--no-transform} should be used only in the rare cases when the
1329instructions must be exactly as specified in the assembly source.
00e3c476
BW
1330
1331@item --rename-section @var{oldname}=@var{newname}
1332When generating output sections, rename the @var{oldname} section to
1333@var{newname}.
e0001a05
NC
1334@end table
1335@end ifset
1336
3c9b82ba
NC
1337@ifset Z80
1338The following options are available when @value{AS} is configured for
1339a Z80 family processor.
1340@table @gcctabopt
1341@item -z80
1342Assemble for Z80 processor.
1343@item -r800
1344Assemble for R800 processor.
01642c12 1345@item -ignore-undocumented-instructions
3c9b82ba
NC
1346@itemx -Wnud
1347Assemble undocumented Z80 instructions that also work on R800 without warning.
01642c12 1348@item -ignore-unportable-instructions
3c9b82ba
NC
1349@itemx -Wnup
1350Assemble all undocumented Z80 instructions without warning.
01642c12 1351@item -warn-undocumented-instructions
3c9b82ba
NC
1352@itemx -Wud
1353Issue a warning for undocumented Z80 instructions that also work on R800.
01642c12 1354@item -warn-unportable-instructions
3c9b82ba 1355@itemx -Wup
01642c12
RM
1356Issue a warning for undocumented Z80 instructions that do not work on R800.
1357@item -forbid-undocumented-instructions
3c9b82ba
NC
1358@itemx -Fud
1359Treat all undocumented instructions as errors.
01642c12 1360@item -forbid-unportable-instructions
3c9b82ba 1361@itemx -Fup
b45619c0 1362Treat undocumented Z80 instructions that do not work on R800 as errors.
3c9b82ba
NC
1363@end table
1364@end ifset
1365
0285c67d
NC
1366@c man end
1367
252b5132
RH
1368@menu
1369* Manual:: Structure of this Manual
1370* GNU Assembler:: The GNU Assembler
1371* Object Formats:: Object File Formats
1372* Command Line:: Command Line
1373* Input Files:: Input Files
1374* Object:: Output (Object) File
1375* Errors:: Error and Warning Messages
1376@end menu
1377
1378@node Manual
1379@section Structure of this Manual
1380
1381@cindex manual, structure and purpose
1382This manual is intended to describe what you need to know to use
a4fb0134 1383@sc{gnu} @command{@value{AS}}. We cover the syntax expected in source files, including
252b5132 1384notation for symbols, constants, and expressions; the directives that
a4fb0134 1385@command{@value{AS}} understands; and of course how to invoke @command{@value{AS}}.
252b5132
RH
1386
1387@ifclear GENERIC
1388We also cover special features in the @value{TARGET}
a4fb0134 1389configuration of @command{@value{AS}}, including assembler directives.
252b5132
RH
1390@end ifclear
1391@ifset GENERIC
1392This manual also describes some of the machine-dependent features of
1393various flavors of the assembler.
1394@end ifset
1395
1396@cindex machine instructions (not covered)
1397On the other hand, this manual is @emph{not} intended as an introduction
1398to programming in assembly language---let alone programming in general!
1399In a similar vein, we make no attempt to introduce the machine
1400architecture; we do @emph{not} describe the instruction set, standard
1401mnemonics, registers or addressing modes that are standard to a
1402particular architecture.
1403@ifset GENERIC
1404You may want to consult the manufacturer's
1405machine architecture manual for this information.
1406@end ifset
1407@ifclear GENERIC
1408@ifset H8/300
1409For information on the H8/300 machine instruction set, see @cite{H8/300
c2dcd04e
NC
1410Series Programming Manual}. For the H8/300H, see @cite{H8/300H Series
1411Programming Manual} (Renesas).
252b5132 1412@end ifset
252b5132 1413@ifset SH
ef230218
JR
1414For information on the Renesas (formerly Hitachi) / SuperH SH machine instruction set,
1415see @cite{SH-Microcomputer User's Manual} (Renesas) or
1416@cite{SH-4 32-bit CPU Core Architecture} (SuperH) and
1417@cite{SuperH (SH) 64-Bit RISC Series} (SuperH).
252b5132
RH
1418@end ifset
1419@ifset Z8000
1420For information on the Z8000 machine instruction set, see @cite{Z8000 CPU Technical Manual}
1421@end ifset
1422@end ifclear
1423
1424@c I think this is premature---doc@cygnus.com, 17jan1991
1425@ignore
1426Throughout this manual, we assume that you are running @dfn{GNU},
1427the portable operating system from the @dfn{Free Software
1428Foundation, Inc.}. This restricts our attention to certain kinds of
1429computer (in particular, the kinds of computers that @sc{gnu} can run on);
1430once this assumption is granted examples and definitions need less
1431qualification.
1432
a4fb0134 1433@command{@value{AS}} is part of a team of programs that turn a high-level
252b5132
RH
1434human-readable series of instructions into a low-level
1435computer-readable series of instructions. Different versions of
a4fb0134 1436@command{@value{AS}} are used for different kinds of computer.
252b5132
RH
1437@end ignore
1438
1439@c There used to be a section "Terminology" here, which defined
1440@c "contents", "byte", "word", and "long". Defining "word" to any
1441@c particular size is confusing when the .word directive may generate 16
1442@c bits on one machine and 32 bits on another; in general, for the user
1443@c version of this manual, none of these terms seem essential to define.
1444@c They were used very little even in the former draft of the manual;
1445@c this draft makes an effort to avoid them (except in names of
1446@c directives).
1447
1448@node GNU Assembler
1449@section The GNU Assembler
1450
0285c67d
NC
1451@c man begin DESCRIPTION
1452
a4fb0134 1453@sc{gnu} @command{as} is really a family of assemblers.
252b5132 1454@ifclear GENERIC
a4fb0134 1455This manual describes @command{@value{AS}}, a member of that family which is
252b5132
RH
1456configured for the @value{TARGET} architectures.
1457@end ifclear
1458If you use (or have used) the @sc{gnu} assembler on one architecture, you
1459should find a fairly similar environment when you use it on another
1460architecture. Each version has much in common with the others,
1461including object file formats, most assembler directives (often called
1462@dfn{pseudo-ops}) and assembler syntax.@refill
1463
1464@cindex purpose of @sc{gnu} assembler
a4fb0134 1465@command{@value{AS}} is primarily intended to assemble the output of the
252b5132 1466@sc{gnu} C compiler @code{@value{GCC}} for use by the linker
a4fb0134 1467@code{@value{LD}}. Nevertheless, we've tried to make @command{@value{AS}}
252b5132
RH
1468assemble correctly everything that other assemblers for the same
1469machine would assemble.
1470@ifset VAX
1471Any exceptions are documented explicitly (@pxref{Machine Dependencies}).
1472@end ifset
1473@ifset M680X0
1474@c This remark should appear in generic version of manual; assumption
1475@c here is that generic version sets M680x0.
a4fb0134 1476This doesn't mean @command{@value{AS}} always uses the same syntax as another
252b5132
RH
1477assembler for the same architecture; for example, we know of several
1478incompatible versions of 680x0 assembly language syntax.
1479@end ifset
1480
0285c67d
NC
1481@c man end
1482
a4fb0134 1483Unlike older assemblers, @command{@value{AS}} is designed to assemble a source
252b5132
RH
1484program in one pass of the source file. This has a subtle impact on the
1485@kbd{.org} directive (@pxref{Org,,@code{.org}}).
1486
1487@node Object Formats
1488@section Object File Formats
1489
1490@cindex object file format
1491The @sc{gnu} assembler can be configured to produce several alternative
1492object file formats. For the most part, this does not affect how you
1493write assembly language programs; but directives for debugging symbols
1494are typically different in different file formats. @xref{Symbol
1495Attributes,,Symbol Attributes}.
1496@ifclear GENERIC
1497@ifclear MULTI-OBJ
c1253627 1498For the @value{TARGET} target, @command{@value{AS}} is configured to produce
252b5132
RH
1499@value{OBJ-NAME} format object files.
1500@end ifclear
1501@c The following should exhaust all configs that set MULTI-OBJ, ideally
252b5132 1502@ifset I960
a4fb0134 1503On the @value{TARGET}, @command{@value{AS}} can be configured to produce either
252b5132
RH
1504@code{b.out} or COFF format object files.
1505@end ifset
1506@ifset HPPA
a4fb0134 1507On the @value{TARGET}, @command{@value{AS}} can be configured to produce either
252b5132
RH
1508SOM or ELF format object files.
1509@end ifset
1510@end ifclear
1511
1512@node Command Line
1513@section Command Line
1514
1515@cindex command line conventions
0285c67d 1516
a4fb0134 1517After the program name @command{@value{AS}}, the command line may contain
252b5132
RH
1518options and file names. Options may appear in any order, and may be
1519before, after, or between file names. The order of file names is
1520significant.
1521
1522@cindex standard input, as input file
1523@kindex --
1524@file{--} (two hyphens) by itself names the standard input file
a4fb0134 1525explicitly, as one of the files for @command{@value{AS}} to assemble.
252b5132
RH
1526
1527@cindex options, command line
1528Except for @samp{--} any command line argument that begins with a
1529hyphen (@samp{-}) is an option. Each option changes the behavior of
a4fb0134 1530@command{@value{AS}}. No option changes the way another option works. An
252b5132
RH
1531option is a @samp{-} followed by one or more letters; the case of
1532the letter is important. All options are optional.
1533
1534Some options expect exactly one file name to follow them. The file
1535name may either immediately follow the option's letter (compatible
1536with older assemblers) or it may be the next command argument (@sc{gnu}
1537standard). These two command lines are equivalent:
1538
1539@smallexample
1540@value{AS} -o my-object-file.o mumble.s
1541@value{AS} -omy-object-file.o mumble.s
1542@end smallexample
1543
1544@node Input Files
1545@section Input Files
1546
1547@cindex input
1548@cindex source program
1549@cindex files, input
1550We use the phrase @dfn{source program}, abbreviated @dfn{source}, to
a4fb0134 1551describe the program input to one run of @command{@value{AS}}. The program may
252b5132
RH
1552be in one or more files; how the source is partitioned into files
1553doesn't change the meaning of the source.
1554
1555@c I added "con" prefix to "catenation" just to prove I can overcome my
1556@c APL training... doc@cygnus.com
1557The source program is a concatenation of the text in all the files, in the
1558order specified.
1559
0285c67d 1560@c man begin DESCRIPTION
a4fb0134 1561Each time you run @command{@value{AS}} it assembles exactly one source
252b5132
RH
1562program. The source program is made up of one or more files.
1563(The standard input is also a file.)
1564
a4fb0134 1565You give @command{@value{AS}} a command line that has zero or more input file
252b5132
RH
1566names. The input files are read (from left file name to right). A
1567command line argument (in any position) that has no special meaning
1568is taken to be an input file name.
1569
a4fb0134
SC
1570If you give @command{@value{AS}} no file names it attempts to read one input file
1571from the @command{@value{AS}} standard input, which is normally your terminal. You
1572may have to type @key{ctl-D} to tell @command{@value{AS}} there is no more program
252b5132
RH
1573to assemble.
1574
1575Use @samp{--} if you need to explicitly name the standard input file
1576in your command line.
1577
a4fb0134 1578If the source is empty, @command{@value{AS}} produces a small, empty object
252b5132
RH
1579file.
1580
0285c67d
NC
1581@c man end
1582
252b5132
RH
1583@subheading Filenames and Line-numbers
1584
1585@cindex input file linenumbers
1586@cindex line numbers, in input files
1587There are two ways of locating a line in the input file (or files) and
1588either may be used in reporting error messages. One way refers to a line
1589number in a physical file; the other refers to a line number in a
1590``logical'' file. @xref{Errors, ,Error and Warning Messages}.
1591
1592@dfn{Physical files} are those files named in the command line given
a4fb0134 1593to @command{@value{AS}}.
252b5132
RH
1594
1595@dfn{Logical files} are simply names declared explicitly by assembler
1596directives; they bear no relation to physical files. Logical file names help
a4fb0134
SC
1597error messages reflect the original source file, when @command{@value{AS}} source
1598is itself synthesized from other files. @command{@value{AS}} understands the
252b5132
RH
1599@samp{#} directives emitted by the @code{@value{GCC}} preprocessor. See also
1600@ref{File,,@code{.file}}.
1601
1602@node Object
1603@section Output (Object) File
1604
1605@cindex object file
1606@cindex output file
1607@kindex a.out
1608@kindex .o
a4fb0134 1609Every time you run @command{@value{AS}} it produces an output file, which is
252b5132
RH
1610your assembly language program translated into numbers. This file
1611is the object file. Its default name is
1612@ifclear BOUT
1613@code{a.out}.
1614@end ifclear
1615@ifset BOUT
1616@ifset GENERIC
01642c12 1617@code{a.out}, or
252b5132 1618@end ifset
a4fb0134 1619@code{b.out} when @command{@value{AS}} is configured for the Intel 80960.
252b5132 1620@end ifset
a4fb0134 1621You can give it another name by using the @option{-o} option. Conventionally,
252b5132
RH
1622object file names end with @file{.o}. The default name is used for historical
1623reasons: older assemblers were capable of assembling self-contained programs
1624directly into a runnable program. (For some formats, this isn't currently
1625possible, but it can be done for the @code{a.out} format.)
1626
1627@cindex linker
1628@kindex ld
1629The object file is meant for input to the linker @code{@value{LD}}. It contains
1630assembled program code, information to help @code{@value{LD}} integrate
1631the assembled program into a runnable file, and (optionally) symbolic
1632information for the debugger.
1633
1634@c link above to some info file(s) like the description of a.out.
1635@c don't forget to describe @sc{gnu} info as well as Unix lossage.
1636
1637@node Errors
1638@section Error and Warning Messages
1639
0285c67d
NC
1640@c man begin DESCRIPTION
1641
a349d9dd 1642@cindex error messages
252b5132
RH
1643@cindex warning messages
1644@cindex messages from assembler
a4fb0134 1645@command{@value{AS}} may write warnings and error messages to the standard error
252b5132 1646file (usually your terminal). This should not happen when a compiler
a4fb0134
SC
1647runs @command{@value{AS}} automatically. Warnings report an assumption made so
1648that @command{@value{AS}} could keep assembling a flawed program; errors report a
252b5132
RH
1649grave problem that stops the assembly.
1650
0285c67d
NC
1651@c man end
1652
252b5132
RH
1653@cindex format of warning messages
1654Warning messages have the format
1655
1656@smallexample
1657file_name:@b{NNN}:Warning Message Text
1658@end smallexample
1659
1660@noindent
1661@cindex line numbers, in warnings/errors
1662(where @b{NNN} is a line number). If a logical file name has been given
1663(@pxref{File,,@code{.file}}) it is used for the filename, otherwise the name of
1664the current input file is used. If a logical line number was given
1665@ifset GENERIC
1666(@pxref{Line,,@code{.line}})
1667@end ifset
252b5132
RH
1668then it is used to calculate the number printed,
1669otherwise the actual line in the current source file is printed. The
1670message text is intended to be self explanatory (in the grand Unix
1671tradition).
1672
1673@cindex format of error messages
1674Error messages have the format
1675@smallexample
1676file_name:@b{NNN}:FATAL:Error Message Text
1677@end smallexample
1678The file name and line number are derived as for warning
1679messages. The actual message text may be rather less explanatory
1680because many of them aren't supposed to happen.
1681
1682@node Invoking
1683@chapter Command-Line Options
1684
1685@cindex options, all versions of assembler
1686This chapter describes command-line options available in @emph{all}
96e9638b
BW
1687versions of the @sc{gnu} assembler; see @ref{Machine Dependencies},
1688for options specific
252b5132 1689@ifclear GENERIC
c1253627 1690to the @value{TARGET} target.
252b5132
RH
1691@end ifclear
1692@ifset GENERIC
1693to particular machine architectures.
1694@end ifset
1695
0285c67d
NC
1696@c man begin DESCRIPTION
1697
c1253627 1698If you are invoking @command{@value{AS}} via the @sc{gnu} C compiler,
252b5132
RH
1699you can use the @samp{-Wa} option to pass arguments through to the assembler.
1700The assembler arguments must be separated from each other (and the @samp{-Wa})
1701by commas. For example:
1702
1703@smallexample
1704gcc -c -g -O -Wa,-alh,-L file.c
1705@end smallexample
1706
1707@noindent
1708This passes two options to the assembler: @samp{-alh} (emit a listing to
5f5e16be 1709standard output with high-level and assembly source) and @samp{-L} (retain
252b5132
RH
1710local symbols in the symbol table).
1711
1712Usually you do not need to use this @samp{-Wa} mechanism, since many compiler
1713command-line options are automatically passed to the assembler by the compiler.
1714(You can call the @sc{gnu} compiler driver with the @samp{-v} option to see
1715precisely what options it passes to each compilation pass, including the
1716assembler.)
1717
0285c67d
NC
1718@c man end
1719
252b5132 1720@menu
83f10cb2 1721* a:: -a[cdghlns] enable listings
caa32fe5 1722* alternate:: --alternate enable alternate macro syntax
252b5132
RH
1723* D:: -D for compatibility
1724* f:: -f to work faster
1725* I:: -I for .include search path
1726@ifclear DIFF-TBL-KLUGE
1727* K:: -K for compatibility
1728@end ifclear
1729@ifset DIFF-TBL-KLUGE
1730* K:: -K for difference tables
1731@end ifset
1732
ba83aca1 1733* L:: -L to retain local symbols
c3a27914 1734* listing:: --listing-XXX to configure listing output
252b5132
RH
1735* M:: -M or --mri to assemble in MRI compatibility mode
1736* MD:: --MD for dependency tracking
1737* o:: -o to name the object file
1738* R:: -R to join data and text sections
1739* statistics:: --statistics to see statistics about assembly
1740* traditional-format:: --traditional-format for compatible output
1741* v:: -v to announce version
2bdd6cf5 1742* W:: -W, --no-warn, --warn, --fatal-warnings to control warnings
252b5132
RH
1743* Z:: -Z to make object file even after errors
1744@end menu
1745
1746@node a
83f10cb2 1747@section Enable Listings: @option{-a[cdghlns]}
252b5132
RH
1748
1749@kindex -a
1750@kindex -ac
1751@kindex -ad
83f10cb2 1752@kindex -ag
252b5132
RH
1753@kindex -ah
1754@kindex -al
1755@kindex -an
1756@kindex -as
1757@cindex listings, enabling
1758@cindex assembly listings, enabling
1759
1760These options enable listing output from the assembler. By itself,
1761@samp{-a} requests high-level, assembly, and symbols listing.
1762You can use other letters to select specific options for the list:
1763@samp{-ah} requests a high-level language listing,
1764@samp{-al} requests an output-program assembly listing, and
1765@samp{-as} requests a symbol table listing.
1766High-level listings require that a compiler debugging option like
1767@samp{-g} be used, and that assembly listings (@samp{-al}) be requested
1768also.
1769
83f10cb2
NC
1770Use the @samp{-ag} option to print a first section with general assembly
1771information, like @value{AS} version, switches passed, or time stamp.
1772
252b5132
RH
1773Use the @samp{-ac} option to omit false conditionals from a listing. Any lines
1774which are not assembled because of a false @code{.if} (or @code{.ifdef}, or any
1775other conditional), or a true @code{.if} followed by an @code{.else}, will be
1776omitted from the listing.
1777
1778Use the @samp{-ad} option to omit debugging directives from the
1779listing.
1780
1781Once you have specified one of these options, you can further control
1782listing output and its appearance using the directives @code{.list},
1783@code{.nolist}, @code{.psize}, @code{.eject}, @code{.title}, and
1784@code{.sbttl}.
1785The @samp{-an} option turns off all forms processing.
1786If you do not request listing output with one of the @samp{-a} options, the
1787listing-control directives have no effect.
1788
1789The letters after @samp{-a} may be combined into one option,
1790@emph{e.g.}, @samp{-aln}.
1791
96e9638b
BW
1792Note if the assembler source is coming from the standard input (e.g.,
1793because it
c3a27914
NC
1794is being created by @code{@value{GCC}} and the @samp{-pipe} command line switch
1795is being used) then the listing will not contain any comments or preprocessor
1796directives. This is because the listing code buffers input source lines from
1797stdin only after they have been preprocessed by the assembler. This reduces
1798memory usage and makes the code more efficient.
1799
caa32fe5
NC
1800@node alternate
1801@section @option{--alternate}
1802
1803@kindex --alternate
1804Begin in alternate macro mode, see @ref{Altmacro,,@code{.altmacro}}.
1805
252b5132 1806@node D
a4fb0134 1807@section @option{-D}
252b5132
RH
1808
1809@kindex -D
1810This option has no effect whatsoever, but it is accepted to make it more
1811likely that scripts written for other assemblers also work with
a4fb0134 1812@command{@value{AS}}.
252b5132
RH
1813
1814@node f
a4fb0134 1815@section Work Faster: @option{-f}
252b5132
RH
1816
1817@kindex -f
1818@cindex trusted compiler
a4fb0134 1819@cindex faster processing (@option{-f})
252b5132
RH
1820@samp{-f} should only be used when assembling programs written by a
1821(trusted) compiler. @samp{-f} stops the assembler from doing whitespace
1822and comment preprocessing on
1823the input file(s) before assembling them. @xref{Preprocessing,
1824,Preprocessing}.
1825
1826@quotation
1827@emph{Warning:} if you use @samp{-f} when the files actually need to be
a4fb0134 1828preprocessed (if they contain comments, for example), @command{@value{AS}} does
252b5132
RH
1829not work correctly.
1830@end quotation
1831
1832@node I
c1253627 1833@section @code{.include} Search Path: @option{-I} @var{path}
252b5132
RH
1834
1835@kindex -I @var{path}
1836@cindex paths for @code{.include}
1837@cindex search path for @code{.include}
1838@cindex @code{include} directive search path
1839Use this option to add a @var{path} to the list of directories
a4fb0134
SC
1840@command{@value{AS}} searches for files specified in @code{.include}
1841directives (@pxref{Include,,@code{.include}}). You may use @option{-I} as
252b5132 1842many times as necessary to include a variety of paths. The current
a4fb0134 1843working directory is always searched first; after that, @command{@value{AS}}
252b5132
RH
1844searches any @samp{-I} directories in the same order as they were
1845specified (left to right) on the command line.
1846
1847@node K
a4fb0134 1848@section Difference Tables: @option{-K}
252b5132
RH
1849
1850@kindex -K
1851@ifclear DIFF-TBL-KLUGE
1852On the @value{TARGET} family, this option is allowed, but has no effect. It is
1853permitted for compatibility with the @sc{gnu} assembler on other platforms,
1854where it can be used to warn when the assembler alters the machine code
1855generated for @samp{.word} directives in difference tables. The @value{TARGET}
1856family does not have the addressing limitations that sometimes lead to this
1857alteration on other platforms.
1858@end ifclear
1859
1860@ifset DIFF-TBL-KLUGE
1861@cindex difference tables, warning
1862@cindex warning for altered difference tables
96e9638b
BW
1863@command{@value{AS}} sometimes alters the code emitted for directives of the
1864form @samp{.word @var{sym1}-@var{sym2}}. @xref{Word,,@code{.word}}.
252b5132
RH
1865You can use the @samp{-K} option if you want a warning issued when this
1866is done.
1867@end ifset
1868
1869@node L
ba83aca1 1870@section Include Local Symbols: @option{-L}
252b5132
RH
1871
1872@kindex -L
ba83aca1
BW
1873@cindex local symbols, retaining in output
1874Symbols beginning with system-specific local label prefixes, typically
1875@samp{.L} for ELF systems or @samp{L} for traditional a.out systems, are
1876called @dfn{local symbols}. @xref{Symbol Names}. Normally you do not see
1877such symbols when debugging, because they are intended for the use of
1878programs (like compilers) that compose assembler programs, not for your
1879notice. Normally both @command{@value{AS}} and @code{@value{LD}} discard
1880such symbols, so you do not normally debug with them.
1881
1882This option tells @command{@value{AS}} to retain those local symbols
252b5132 1883in the object file. Usually if you do this you also tell the linker
ba83aca1 1884@code{@value{LD}} to preserve those symbols.
252b5132 1885
c3a27914 1886@node listing
a4fb0134 1887@section Configuring listing output: @option{--listing}
c3a27914
NC
1888
1889The listing feature of the assembler can be enabled via the command line switch
1890@samp{-a} (@pxref{a}). This feature combines the input source file(s) with a
1891hex dump of the corresponding locations in the output object file, and displays
96e9638b
BW
1892them as a listing file. The format of this listing can be controlled by
1893directives inside the assembler source (i.e., @code{.list} (@pxref{List}),
1894@code{.title} (@pxref{Title}), @code{.sbttl} (@pxref{Sbttl}),
1895@code{.psize} (@pxref{Psize}), and
1896@code{.eject} (@pxref{Eject}) and also by the following switches:
c3a27914 1897
a4fb0134 1898@table @gcctabopt
c3a27914
NC
1899@item --listing-lhs-width=@samp{number}
1900@kindex --listing-lhs-width
1901@cindex Width of first line disassembly output
1902Sets the maximum width, in words, of the first line of the hex byte dump. This
1903dump appears on the left hand side of the listing output.
1904
1905@item --listing-lhs-width2=@samp{number}
1906@kindex --listing-lhs-width2
1907@cindex Width of continuation lines of disassembly output
1908Sets the maximum width, in words, of any further lines of the hex byte dump for
8dfa0188 1909a given input source line. If this value is not specified, it defaults to being
c3a27914
NC
1910the same as the value specified for @samp{--listing-lhs-width}. If neither
1911switch is used the default is to one.
1912
1913@item --listing-rhs-width=@samp{number}
1914@kindex --listing-rhs-width
1915@cindex Width of source line output
1916Sets the maximum width, in characters, of the source line that is displayed
1917alongside the hex dump. The default value for this parameter is 100. The
1918source line is displayed on the right hand side of the listing output.
1919
1920@item --listing-cont-lines=@samp{number}
1921@kindex --listing-cont-lines
1922@cindex Maximum number of continuation lines
1923Sets the maximum number of continuation lines of hex dump that will be
1924displayed for a given single line of source input. The default value is 4.
1925@end table
1926
252b5132 1927@node M
a4fb0134 1928@section Assemble in MRI Compatibility Mode: @option{-M}
252b5132
RH
1929
1930@kindex -M
1931@cindex MRI compatibility mode
a4fb0134
SC
1932The @option{-M} or @option{--mri} option selects MRI compatibility mode. This
1933changes the syntax and pseudo-op handling of @command{@value{AS}} to make it
252b5132
RH
1934compatible with the @code{ASM68K} or the @code{ASM960} (depending upon the
1935configured target) assembler from Microtec Research. The exact nature of the
1936MRI syntax will not be documented here; see the MRI manuals for more
1937information. Note in particular that the handling of macros and macro
1938arguments is somewhat different. The purpose of this option is to permit
a4fb0134 1939assembling existing MRI assembler code using @command{@value{AS}}.
252b5132
RH
1940
1941The MRI compatibility is not complete. Certain operations of the MRI assembler
1942depend upon its object file format, and can not be supported using other object
1943file formats. Supporting these would require enhancing each object file format
1944individually. These are:
1945
1946@itemize @bullet
1947@item global symbols in common section
1948
1949The m68k MRI assembler supports common sections which are merged by the linker.
a4fb0134 1950Other object file formats do not support this. @command{@value{AS}} handles
252b5132
RH
1951common sections by treating them as a single common symbol. It permits local
1952symbols to be defined within a common section, but it can not support global
1953symbols, since it has no way to describe them.
1954
1955@item complex relocations
1956
1957The MRI assemblers support relocations against a negated section address, and
1958relocations which combine the start addresses of two or more sections. These
1959are not support by other object file formats.
1960
1961@item @code{END} pseudo-op specifying start address
1962
1963The MRI @code{END} pseudo-op permits the specification of a start address.
1964This is not supported by other object file formats. The start address may
a4fb0134 1965instead be specified using the @option{-e} option to the linker, or in a linker
252b5132
RH
1966script.
1967
1968@item @code{IDNT}, @code{.ident} and @code{NAME} pseudo-ops
1969
1970The MRI @code{IDNT}, @code{.ident} and @code{NAME} pseudo-ops assign a module
1971name to the output file. This is not supported by other object file formats.
1972
1973@item @code{ORG} pseudo-op
1974
1975The m68k MRI @code{ORG} pseudo-op begins an absolute section at a given
a4fb0134 1976address. This differs from the usual @command{@value{AS}} @code{.org} pseudo-op,
252b5132
RH
1977which changes the location within the current section. Absolute sections are
1978not supported by other object file formats. The address of a section may be
1979assigned within a linker script.
1980@end itemize
1981
1982There are some other features of the MRI assembler which are not supported by
a4fb0134 1983@command{@value{AS}}, typically either because they are difficult or because they
252b5132
RH
1984seem of little consequence. Some of these may be supported in future releases.
1985
1986@itemize @bullet
1987
1988@item EBCDIC strings
1989
1990EBCDIC strings are not supported.
1991
1992@item packed binary coded decimal
1993
1994Packed binary coded decimal is not supported. This means that the @code{DC.P}
1995and @code{DCB.P} pseudo-ops are not supported.
1996
1997@item @code{FEQU} pseudo-op
1998
1999The m68k @code{FEQU} pseudo-op is not supported.
2000
2001@item @code{NOOBJ} pseudo-op
2002
2003The m68k @code{NOOBJ} pseudo-op is not supported.
2004
2005@item @code{OPT} branch control options
2006
2007The m68k @code{OPT} branch control options---@code{B}, @code{BRS}, @code{BRB},
a4fb0134 2008@code{BRL}, and @code{BRW}---are ignored. @command{@value{AS}} automatically
252b5132
RH
2009relaxes all branches, whether forward or backward, to an appropriate size, so
2010these options serve no purpose.
2011
2012@item @code{OPT} list control options
2013
2014The following m68k @code{OPT} list control options are ignored: @code{C},
2015@code{CEX}, @code{CL}, @code{CRE}, @code{E}, @code{G}, @code{I}, @code{M},
2016@code{MEX}, @code{MC}, @code{MD}, @code{X}.
2017
2018@item other @code{OPT} options
2019
2020The following m68k @code{OPT} options are ignored: @code{NEST}, @code{O},
2021@code{OLD}, @code{OP}, @code{P}, @code{PCO}, @code{PCR}, @code{PCS}, @code{R}.
2022
2023@item @code{OPT} @code{D} option is default
2024
2025The m68k @code{OPT} @code{D} option is the default, unlike the MRI assembler.
2026@code{OPT NOD} may be used to turn it off.
2027
2028@item @code{XREF} pseudo-op.
2029
2030The m68k @code{XREF} pseudo-op is ignored.
2031
2032@item @code{.debug} pseudo-op
2033
2034The i960 @code{.debug} pseudo-op is not supported.
2035
2036@item @code{.extended} pseudo-op
2037
2038The i960 @code{.extended} pseudo-op is not supported.
2039
2040@item @code{.list} pseudo-op.
2041
2042The various options of the i960 @code{.list} pseudo-op are not supported.
2043
2044@item @code{.optimize} pseudo-op
2045
2046The i960 @code{.optimize} pseudo-op is not supported.
2047
2048@item @code{.output} pseudo-op
2049
2050The i960 @code{.output} pseudo-op is not supported.
2051
2052@item @code{.setreal} pseudo-op
2053
2054The i960 @code{.setreal} pseudo-op is not supported.
2055
2056@end itemize
2057
2058@node MD
c1253627 2059@section Dependency Tracking: @option{--MD}
252b5132
RH
2060
2061@kindex --MD
2062@cindex dependency tracking
2063@cindex make rules
2064
a4fb0134 2065@command{@value{AS}} can generate a dependency file for the file it creates. This
252b5132
RH
2066file consists of a single rule suitable for @code{make} describing the
2067dependencies of the main source file.
2068
2069The rule is written to the file named in its argument.
2070
2071This feature is used in the automatic updating of makefiles.
2072
2073@node o
a4fb0134 2074@section Name the Object File: @option{-o}
252b5132
RH
2075
2076@kindex -o
2077@cindex naming object file
2078@cindex object file name
a4fb0134 2079There is always one object file output when you run @command{@value{AS}}. By
252b5132
RH
2080default it has the name
2081@ifset GENERIC
2082@ifset I960
2083@file{a.out} (or @file{b.out}, for Intel 960 targets only).
2084@end ifset
2085@ifclear I960
2086@file{a.out}.
2087@end ifclear
2088@end ifset
2089@ifclear GENERIC
2090@ifset I960
2091@file{b.out}.
2092@end ifset
2093@ifclear I960
2094@file{a.out}.
2095@end ifclear
2096@end ifclear
2097You use this option (which takes exactly one filename) to give the
2098object file a different name.
2099
a4fb0134 2100Whatever the object file is called, @command{@value{AS}} overwrites any
252b5132
RH
2101existing file of the same name.
2102
2103@node R
a4fb0134 2104@section Join Data and Text Sections: @option{-R}
252b5132
RH
2105
2106@kindex -R
2107@cindex data and text sections, joining
2108@cindex text and data sections, joining
2109@cindex joining text and data sections
2110@cindex merging text and data sections
a4fb0134 2111@option{-R} tells @command{@value{AS}} to write the object file as if all
252b5132
RH
2112data-section data lives in the text section. This is only done at
2113the very last moment: your binary data are the same, but data
2114section parts are relocated differently. The data section part of
2115your object file is zero bytes long because all its bytes are
2116appended to the text section. (@xref{Sections,,Sections and Relocation}.)
2117
a4fb0134 2118When you specify @option{-R} it would be possible to generate shorter
252b5132
RH
2119address displacements (because we do not have to cross between text and
2120data section). We refrain from doing this simply for compatibility with
a4fb0134 2121older versions of @command{@value{AS}}. In future, @option{-R} may work this way.
252b5132 2122
c1253627
NC
2123@ifset COFF-ELF
2124When @command{@value{AS}} is configured for COFF or ELF output,
252b5132
RH
2125this option is only useful if you use sections named @samp{.text} and
2126@samp{.data}.
2127@end ifset
2128
2129@ifset HPPA
a4fb0134
SC
2130@option{-R} is not supported for any of the HPPA targets. Using
2131@option{-R} generates a warning from @command{@value{AS}}.
252b5132
RH
2132@end ifset
2133
2134@node statistics
a4fb0134 2135@section Display Assembly Statistics: @option{--statistics}
252b5132
RH
2136
2137@kindex --statistics
2138@cindex statistics, about assembly
2139@cindex time, total for assembly
2140@cindex space used, maximum for assembly
2141Use @samp{--statistics} to display two statistics about the resources used by
a4fb0134 2142@command{@value{AS}}: the maximum amount of space allocated during the assembly
252b5132
RH
2143(in bytes), and the total execution time taken for the assembly (in @sc{cpu}
2144seconds).
2145
2146@node traditional-format
c1253627 2147@section Compatible Output: @option{--traditional-format}
252b5132
RH
2148
2149@kindex --traditional-format
a4fb0134 2150For some targets, the output of @command{@value{AS}} is different in some ways
252b5132 2151from the output of some existing assembler. This switch requests
a4fb0134 2152@command{@value{AS}} to use the traditional format instead.
252b5132
RH
2153
2154For example, it disables the exception frame optimizations which
a4fb0134 2155@command{@value{AS}} normally does by default on @code{@value{GCC}} output.
252b5132
RH
2156
2157@node v
a4fb0134 2158@section Announce Version: @option{-v}
252b5132
RH
2159
2160@kindex -v
2161@kindex -version
2162@cindex assembler version
2163@cindex version of assembler
2164You can find out what version of as is running by including the
2165option @samp{-v} (which you can also spell as @samp{-version}) on the
2166command line.
2167
2168@node W
a4fb0134 2169@section Control Warnings: @option{-W}, @option{--warn}, @option{--no-warn}, @option{--fatal-warnings}
252b5132 2170
a4fb0134 2171@command{@value{AS}} should never give a warning or error message when
252b5132 2172assembling compiler output. But programs written by people often
a4fb0134 2173cause @command{@value{AS}} to give a warning that a particular assumption was
252b5132 2174made. All such warnings are directed to the standard error file.
2bdd6cf5 2175
c1253627
NC
2176@kindex -W
2177@kindex --no-warn
2bdd6cf5
GK
2178@cindex suppressing warnings
2179@cindex warnings, suppressing
a4fb0134 2180If you use the @option{-W} and @option{--no-warn} options, no warnings are issued.
2bdd6cf5 2181This only affects the warning messages: it does not change any particular of
a4fb0134 2182how @command{@value{AS}} assembles your file. Errors, which stop the assembly,
2bdd6cf5
GK
2183are still reported.
2184
c1253627 2185@kindex --fatal-warnings
2bdd6cf5
GK
2186@cindex errors, caused by warnings
2187@cindex warnings, causing error
a4fb0134 2188If you use the @option{--fatal-warnings} option, @command{@value{AS}} considers
2bdd6cf5
GK
2189files that generate warnings to be in error.
2190
c1253627 2191@kindex --warn
2bdd6cf5 2192@cindex warnings, switching on
a4fb0134 2193You can switch these options off again by specifying @option{--warn}, which
2bdd6cf5 2194causes warnings to be output as usual.
252b5132
RH
2195
2196@node Z
a4fb0134 2197@section Generate Object File in Spite of Errors: @option{-Z}
252b5132
RH
2198@cindex object file, after errors
2199@cindex errors, continuing after
a4fb0134 2200After an error message, @command{@value{AS}} normally produces no output. If for
252b5132 2201some reason you are interested in object file output even after
a4fb0134
SC
2202@command{@value{AS}} gives an error message on your program, use the @samp{-Z}
2203option. If there are any errors, @command{@value{AS}} continues anyways, and
252b5132
RH
2204writes an object file after a final warning message of the form @samp{@var{n}
2205errors, @var{m} warnings, generating bad object file.}
2206
2207@node Syntax
2208@chapter Syntax
2209
2210@cindex machine-independent syntax
2211@cindex syntax, machine-independent
2212This chapter describes the machine-independent syntax allowed in a
a4fb0134 2213source file. @command{@value{AS}} syntax is similar to what many other
252b5132
RH
2214assemblers use; it is inspired by the BSD 4.2
2215@ifclear VAX
2216assembler.
2217@end ifclear
2218@ifset VAX
a4fb0134 2219assembler, except that @command{@value{AS}} does not assemble Vax bit-fields.
252b5132
RH
2220@end ifset
2221
2222@menu
2223* Preprocessing:: Preprocessing
2224* Whitespace:: Whitespace
2225* Comments:: Comments
2226* Symbol Intro:: Symbols
2227* Statements:: Statements
2228* Constants:: Constants
2229@end menu
2230
2231@node Preprocessing
2232@section Preprocessing
2233
2234@cindex preprocessing
a4fb0134 2235The @command{@value{AS}} internal preprocessor:
252b5132
RH
2236@itemize @bullet
2237@cindex whitespace, removed by preprocessor
2238@item
2239adjusts and removes extra whitespace. It leaves one space or tab before
2240the keywords on a line, and turns any other whitespace on the line into
2241a single space.
2242
2243@cindex comments, removed by preprocessor
2244@item
2245removes all comments, replacing them with a single space, or an
2246appropriate number of newlines.
2247
2248@cindex constants, converted by preprocessor
2249@item
2250converts character constants into the appropriate numeric values.
2251@end itemize
2252
2253It does not do macro processing, include file handling, or
2254anything else you may get from your C compiler's preprocessor. You can
2255do include file processing with the @code{.include} directive
2256(@pxref{Include,,@code{.include}}). You can use the @sc{gnu} C compiler driver
c1253627 2257to get other ``CPP'' style preprocessing by giving the input file a
96e9638b 2258@samp{.S} suffix. @xref{Overall Options, ,Options Controlling the Kind of
252b5132
RH
2259Output, gcc.info, Using GNU CC}.
2260
2261Excess whitespace, comments, and character constants
2262cannot be used in the portions of the input text that are not
2263preprocessed.
2264
2265@cindex turning preprocessing on and off
2266@cindex preprocessing, turning on and off
2267@kindex #NO_APP
2268@kindex #APP
2269If the first line of an input file is @code{#NO_APP} or if you use the
2270@samp{-f} option, whitespace and comments are not removed from the input file.
2271Within an input file, you can ask for whitespace and comment removal in
2272specific portions of the by putting a line that says @code{#APP} before the
2273text that may contain whitespace or comments, and putting a line that says
2274@code{#NO_APP} after this text. This feature is mainly intend to support
2275@code{asm} statements in compilers whose output is otherwise free of comments
2276and whitespace.
2277
2278@node Whitespace
2279@section Whitespace
2280
2281@cindex whitespace
2282@dfn{Whitespace} is one or more blanks or tabs, in any order.
2283Whitespace is used to separate symbols, and to make programs neater for
2284people to read. Unless within character constants
2285(@pxref{Characters,,Character Constants}), any whitespace means the same
2286as exactly one space.
2287
2288@node Comments
2289@section Comments
2290
2291@cindex comments
a4fb0134 2292There are two ways of rendering comments to @command{@value{AS}}. In both
252b5132
RH
2293cases the comment is equivalent to one space.
2294
2295Anything from @samp{/*} through the next @samp{*/} is a comment.
2296This means you may not nest these comments.
2297
2298@smallexample
2299/*
2300 The only way to include a newline ('\n') in a comment
2301 is to use this sort of comment.
2302*/
2303
2304/* This sort of comment does not nest. */
2305@end smallexample
2306
2307@cindex line comment character
2308Anything from the @dfn{line comment} character to the next newline
2309is considered a comment and is ignored. The line comment character is
252b5132
RH
2310@ifset ARC
2311@samp{;} on the ARC;
2312@end ifset
550262c4
NC
2313@ifset ARM
2314@samp{@@} on the ARM;
2315@end ifset
252b5132
RH
2316@ifset H8/300
2317@samp{;} for the H8/300 family;
2318@end ifset
252b5132
RH
2319@ifset HPPA
2320@samp{;} for the HPPA;
2321@end ifset
55b62671
AJ
2322@ifset I80386
2323@samp{#} on the i386 and x86-64;
2324@end ifset
252b5132
RH
2325@ifset I960
2326@samp{#} on the i960;
2327@end ifset
e135f41b
NC
2328@ifset PDP11
2329@samp{;} for the PDP-11;
2330@end ifset
041dd5a9
ILT
2331@ifset PJ
2332@samp{;} for picoJava;
2333@end ifset
418c1742 2334@ifset PPC
3fb9d77f 2335@samp{#} for Motorola PowerPC;
418c1742 2336@end ifset
11c19e16
MS
2337@ifset S390
2338@samp{#} for IBM S/390;
2339@end ifset
c0157db4
NC
2340@ifset SCORE
2341@samp{#} for the Sunplus SCORE;
2342@end ifset
252b5132 2343@ifset SH
ef230218 2344@samp{!} for the Renesas / SuperH SH;
252b5132
RH
2345@end ifset
2346@ifset SPARC
2347@samp{!} on the SPARC;
2348@end ifset
a40cbfa3
NC
2349@ifset IP2K
2350@samp{#} on the ip2k;
2351@end ifset
49f58d10
JB
2352@ifset M32C
2353@samp{#} on the m32c;
2354@end ifset
252b5132
RH
2355@ifset M32R
2356@samp{#} on the m32r;
2357@end ifset
2358@ifset M680X0
2359@samp{|} on the 680x0;
2360@end ifset
60bcf0fa
NC
2361@ifset M68HC11
2362@samp{#} on the 68HC11 and 68HC12;
2363@end ifset
046d31c2
NC
2364@ifset RX
2365@samp{#} on the RX;
2366@end ifset
40b36596
JM
2367@ifset TIC6X
2368@samp{;} on the TMS320C6X;
2369@end ifset
252b5132
RH
2370@ifset VAX
2371@samp{#} on the Vax;
2372@end ifset
3c9b82ba
NC
2373@ifset Z80
2374@samp{;} for the Z80;
2375@end ifset
252b5132
RH
2376@ifset Z8000
2377@samp{!} for the Z8000;
2378@end ifset
2379@ifset V850
2380@samp{#} on the V850;
2381@end ifset
e0001a05
NC
2382@ifset XTENSA
2383@samp{#} for Xtensa systems;
2384@end ifset
252b5132 2385see @ref{Machine Dependencies}. @refill
81b0b3f1 2386@c FIXME What about i860?
252b5132
RH
2387
2388@ifset GENERIC
2389On some machines there are two different line comment characters. One
2390character only begins a comment if it is the first non-whitespace character on
2391a line, while the other always begins a comment.
2392@end ifset
2393
2394@ifset V850
2395The V850 assembler also supports a double dash as starting a comment that
2396extends to the end of the line.
2397
2398@samp{--};
2399@end ifset
2400
2401@kindex #
2402@cindex lines starting with @code{#}
2403@cindex logical line numbers
2404To be compatible with past assemblers, lines that begin with @samp{#} have a
2405special interpretation. Following the @samp{#} should be an absolute
2406expression (@pxref{Expressions}): the logical line number of the @emph{next}
96e9638b 2407line. Then a string (@pxref{Strings, ,Strings}) is allowed: if present it is a
252b5132
RH
2408new logical file name. The rest of the line, if any, should be whitespace.
2409
2410If the first non-whitespace characters on the line are not numeric,
2411the line is ignored. (Just like a comment.)
2412
2413@smallexample
2414 # This is an ordinary comment.
2415# 42-6 "new_file_name" # New logical file name
2416 # This is logical line # 36.
2417@end smallexample
2418This feature is deprecated, and may disappear from future versions
a4fb0134 2419of @command{@value{AS}}.
252b5132
RH
2420
2421@node Symbol Intro
2422@section Symbols
2423
2424@cindex characters used in symbols
2425@ifclear SPECIAL-SYMS
2426A @dfn{symbol} is one or more characters chosen from the set of all
2427letters (both upper and lower case), digits and the three characters
2428@samp{_.$}.
2429@end ifclear
2430@ifset SPECIAL-SYMS
2431@ifclear GENERIC
2432@ifset H8
2433A @dfn{symbol} is one or more characters chosen from the set of all
2434letters (both upper and lower case), digits and the three characters
2435@samp{._$}. (Save that, on the H8/300 only, you may not use @samp{$} in
2436symbol names.)
2437@end ifset
2438@end ifclear
2439@end ifset
2440@ifset GENERIC
2441On most machines, you can also use @code{$} in symbol names; exceptions
2442are noted in @ref{Machine Dependencies}.
2443@end ifset
2444No symbol may begin with a digit. Case is significant.
2445There is no length limit: all characters are significant. Symbols are
2446delimited by characters not in that set, or by the beginning of a file
2447(since the source program must end with a newline, the end of a file is
2448not a possible symbol delimiter). @xref{Symbols}.
2449@cindex length of symbols
2450
2451@node Statements
2452@section Statements
2453
2454@cindex statements, structure of
2455@cindex line separator character
2456@cindex statement separator character
2457@ifclear GENERIC
2458@ifclear abnormal-separator
2459A @dfn{statement} ends at a newline character (@samp{\n}) or at a
2460semicolon (@samp{;}). The newline or semicolon is considered part of
2461the preceding statement. Newlines and semicolons within character
2462constants are an exception: they do not end statements.
2463@end ifclear
2464@ifset abnormal-separator
252b5132 2465@ifset HPPA
01642c12 2466A @dfn{statement} ends at a newline character (@samp{\n}) or an exclamation
252b5132
RH
2467point (@samp{!}). The newline or exclamation point is considered part of the
2468preceding statement. Newlines and exclamation points within character
2469constants are an exception: they do not end statements.
2470@end ifset
2471@ifset H8
2472A @dfn{statement} ends at a newline character (@samp{\n}); or (for the
7be1c489 2473H8/300) a dollar sign (@samp{$}); or (for the Renesas-SH) a semicolon
252b5132
RH
2474(@samp{;}). The newline or separator character is considered part of
2475the preceding statement. Newlines and separators within character
2476constants are an exception: they do not end statements.
2477@end ifset
2478@end ifset
2479@end ifclear
2480@ifset GENERIC
2481A @dfn{statement} ends at a newline character (@samp{\n}) or line
96e9638b
BW
2482separator character. (The line separator is usually @samp{;}, unless this
2483conflicts with the comment character; see @ref{Machine Dependencies}.) The
252b5132
RH
2484newline or separator character is considered part of the preceding
2485statement. Newlines and separators within character constants are an
2486exception: they do not end statements.
2487@end ifset
2488
2489@cindex newline, required at file end
2490@cindex EOF, newline must precede
2491It is an error to end any statement with end-of-file: the last
2492character of any input file should be a newline.@refill
2493
2494An empty statement is allowed, and may include whitespace. It is ignored.
2495
2496@cindex instructions and directives
2497@cindex directives and instructions
2498@c "key symbol" is not used elsewhere in the document; seems pedantic to
2499@c @defn{} it in that case, as was done previously... doc@cygnus.com,
2500@c 13feb91.
2501A statement begins with zero or more labels, optionally followed by a
2502key symbol which determines what kind of statement it is. The key
2503symbol determines the syntax of the rest of the statement. If the
2504symbol begins with a dot @samp{.} then the statement is an assembler
2505directive: typically valid for any computer. If the symbol begins with
2506a letter the statement is an assembly language @dfn{instruction}: it
2507assembles into a machine language instruction.
2508@ifset GENERIC
a4fb0134 2509Different versions of @command{@value{AS}} for different computers
252b5132
RH
2510recognize different instructions. In fact, the same symbol may
2511represent a different instruction in a different computer's assembly
2512language.@refill
2513@end ifset
2514
2515@cindex @code{:} (label)
2516@cindex label (@code{:})
2517A label is a symbol immediately followed by a colon (@code{:}).
2518Whitespace before a label or after a colon is permitted, but you may not
2519have whitespace between a label's symbol and its colon. @xref{Labels}.
2520
2521@ifset HPPA
01642c12 2522For HPPA targets, labels need not be immediately followed by a colon, but
252b5132
RH
2523the definition of a label must begin in column zero. This also implies that
2524only one label may be defined on each line.
2525@end ifset
2526
2527@smallexample
2528label: .directive followed by something
2529another_label: # This is an empty statement.
2530 instruction operand_1, operand_2, @dots{}
2531@end smallexample
2532
2533@node Constants
2534@section Constants
2535
2536@cindex constants
2537A constant is a number, written so that its value is known by
2538inspection, without knowing any context. Like this:
2539@smallexample
2540@group
2541.byte 74, 0112, 092, 0x4A, 0X4a, 'J, '\J # All the same value.
2542.ascii "Ring the bell\7" # A string constant.
2543.octa 0x123456789abcdef0123456789ABCDEF0 # A bignum.
2544.float 0f-314159265358979323846264338327\
254595028841971.693993751E-40 # - pi, a flonum.
2546@end group
2547@end smallexample
2548
2549@menu
2550* Characters:: Character Constants
2551* Numbers:: Number Constants
2552@end menu
2553
2554@node Characters
2555@subsection Character Constants
2556
2557@cindex character constants
2558@cindex constants, character
2559There are two kinds of character constants. A @dfn{character} stands
2560for one character in one byte and its value may be used in
2561numeric expressions. String constants (properly called string
2562@emph{literals}) are potentially many bytes and their values may not be
2563used in arithmetic expressions.
2564
2565@menu
2566* Strings:: Strings
2567* Chars:: Characters
2568@end menu
2569
2570@node Strings
2571@subsubsection Strings
2572
2573@cindex string constants
2574@cindex constants, string
2575A @dfn{string} is written between double-quotes. It may contain
2576double-quotes or null characters. The way to get special characters
2577into a string is to @dfn{escape} these characters: precede them with
2578a backslash @samp{\} character. For example @samp{\\} represents
2579one backslash: the first @code{\} is an escape which tells
a4fb0134
SC
2580@command{@value{AS}} to interpret the second character literally as a backslash
2581(which prevents @command{@value{AS}} from recognizing the second @code{\} as an
252b5132
RH
2582escape character). The complete list of escapes follows.
2583
2584@cindex escape codes, character
2585@cindex character escape codes
2586@table @kbd
2587@c @item \a
2588@c Mnemonic for ACKnowledge; for ASCII this is octal code 007.
2589@c
2590@cindex @code{\b} (backspace character)
2591@cindex backspace (@code{\b})
2592@item \b
2593Mnemonic for backspace; for ASCII this is octal code 010.
2594
2595@c @item \e
2596@c Mnemonic for EOText; for ASCII this is octal code 004.
2597@c
2598@cindex @code{\f} (formfeed character)
2599@cindex formfeed (@code{\f})
2600@item \f
2601Mnemonic for FormFeed; for ASCII this is octal code 014.
2602
2603@cindex @code{\n} (newline character)
2604@cindex newline (@code{\n})
2605@item \n
2606Mnemonic for newline; for ASCII this is octal code 012.
2607
2608@c @item \p
2609@c Mnemonic for prefix; for ASCII this is octal code 033, usually known as @code{escape}.
2610@c
2611@cindex @code{\r} (carriage return character)
2612@cindex carriage return (@code{\r})
2613@item \r
2614Mnemonic for carriage-Return; for ASCII this is octal code 015.
2615
2616@c @item \s
2617@c Mnemonic for space; for ASCII this is octal code 040. Included for compliance with
2618@c other assemblers.
2619@c
2620@cindex @code{\t} (tab)
2621@cindex tab (@code{\t})
2622@item \t
2623Mnemonic for horizontal Tab; for ASCII this is octal code 011.
2624
2625@c @item \v
2626@c Mnemonic for Vertical tab; for ASCII this is octal code 013.
2627@c @item \x @var{digit} @var{digit} @var{digit}
2628@c A hexadecimal character code. The numeric code is 3 hexadecimal digits.
2629@c
2630@cindex @code{\@var{ddd}} (octal character code)
2631@cindex octal character code (@code{\@var{ddd}})
2632@item \ @var{digit} @var{digit} @var{digit}
2633An octal character code. The numeric code is 3 octal digits.
2634For compatibility with other Unix systems, 8 and 9 are accepted as digits:
2635for example, @code{\008} has the value 010, and @code{\009} the value 011.
2636
2637@cindex @code{\@var{xd...}} (hex character code)
2638@cindex hex character code (@code{\@var{xd...}})
2639@item \@code{x} @var{hex-digits...}
2640A hex character code. All trailing hex digits are combined. Either upper or
2641lower case @code{x} works.
2642
2643@cindex @code{\\} (@samp{\} character)
2644@cindex backslash (@code{\\})
2645@item \\
2646Represents one @samp{\} character.
2647
2648@c @item \'
2649@c Represents one @samp{'} (accent acute) character.
2650@c This is needed in single character literals
2651@c (@xref{Characters,,Character Constants}.) to represent
2652@c a @samp{'}.
2653@c
2654@cindex @code{\"} (doublequote character)
2655@cindex doublequote (@code{\"})
2656@item \"
2657Represents one @samp{"} character. Needed in strings to represent
2658this character, because an unescaped @samp{"} would end the string.
2659
2660@item \ @var{anything-else}
2661Any other character when escaped by @kbd{\} gives a warning, but
2662assembles as if the @samp{\} was not present. The idea is that if
2663you used an escape sequence you clearly didn't want the literal
a4fb0134
SC
2664interpretation of the following character. However @command{@value{AS}} has no
2665other interpretation, so @command{@value{AS}} knows it is giving you the wrong
252b5132
RH
2666code and warns you of the fact.
2667@end table
2668
2669Which characters are escapable, and what those escapes represent,
2670varies widely among assemblers. The current set is what we think
2671the BSD 4.2 assembler recognizes, and is a subset of what most C
2672compilers recognize. If you are in doubt, do not use an escape
2673sequence.
2674
2675@node Chars
2676@subsubsection Characters
2677
2678@cindex single character constant
2679@cindex character, single
2680@cindex constant, single character
2681A single character may be written as a single quote immediately
2682followed by that character. The same escapes apply to characters as
2683to strings. So if you want to write the character backslash, you
2684must write @kbd{'\\} where the first @code{\} escapes the second
2685@code{\}. As you can see, the quote is an acute accent, not a
2686grave accent. A newline
2687@ifclear GENERIC
2688@ifclear abnormal-separator
2689(or semicolon @samp{;})
2690@end ifclear
2691@ifset abnormal-separator
252b5132
RH
2692@ifset H8
2693(or dollar sign @samp{$}, for the H8/300; or semicolon @samp{;} for the
7be1c489 2694Renesas SH)
252b5132
RH
2695@end ifset
2696@end ifset
2697@end ifclear
2698immediately following an acute accent is taken as a literal character
2699and does not count as the end of a statement. The value of a character
2700constant in a numeric expression is the machine's byte-wide code for
a4fb0134 2701that character. @command{@value{AS}} assumes your character code is ASCII:
252b5132
RH
2702@kbd{'A} means 65, @kbd{'B} means 66, and so on. @refill
2703
2704@node Numbers
2705@subsection Number Constants
2706
2707@cindex constants, number
2708@cindex number constants
a4fb0134 2709@command{@value{AS}} distinguishes three kinds of numbers according to how they
252b5132
RH
2710are stored in the target machine. @emph{Integers} are numbers that
2711would fit into an @code{int} in the C language. @emph{Bignums} are
2712integers, but they are stored in more than 32 bits. @emph{Flonums}
2713are floating point numbers, described below.
2714
2715@menu
2716* Integers:: Integers
2717* Bignums:: Bignums
2718* Flonums:: Flonums
2719@ifclear GENERIC
2720@ifset I960
2721* Bit Fields:: Bit Fields
2722@end ifset
2723@end ifclear
2724@end menu
2725
2726@node Integers
2727@subsubsection Integers
2728@cindex integers
2729@cindex constants, integer
2730
2731@cindex binary integers
2732@cindex integers, binary
2733A binary integer is @samp{0b} or @samp{0B} followed by zero or more of
2734the binary digits @samp{01}.
2735
2736@cindex octal integers
2737@cindex integers, octal
2738An octal integer is @samp{0} followed by zero or more of the octal
2739digits (@samp{01234567}).
2740
2741@cindex decimal integers
2742@cindex integers, decimal
2743A decimal integer starts with a non-zero digit followed by zero or
2744more digits (@samp{0123456789}).
2745
2746@cindex hexadecimal integers
2747@cindex integers, hexadecimal
2748A hexadecimal integer is @samp{0x} or @samp{0X} followed by one or
2749more hexadecimal digits chosen from @samp{0123456789abcdefABCDEF}.
2750
2751Integers have the usual values. To denote a negative integer, use
2752the prefix operator @samp{-} discussed under expressions
2753(@pxref{Prefix Ops,,Prefix Operators}).
2754
2755@node Bignums
2756@subsubsection Bignums
2757
2758@cindex bignums
2759@cindex constants, bignum
2760A @dfn{bignum} has the same syntax and semantics as an integer
2761except that the number (or its negative) takes more than 32 bits to
2762represent in binary. The distinction is made because in some places
2763integers are permitted while bignums are not.
2764
2765@node Flonums
2766@subsubsection Flonums
2767@cindex flonums
2768@cindex floating point numbers
2769@cindex constants, floating point
2770
2771@cindex precision, floating point
2772A @dfn{flonum} represents a floating point number. The translation is
2773indirect: a decimal floating point number from the text is converted by
a4fb0134 2774@command{@value{AS}} to a generic binary floating point number of more than
252b5132
RH
2775sufficient precision. This generic floating point number is converted
2776to a particular computer's floating point format (or formats) by a
a4fb0134 2777portion of @command{@value{AS}} specialized to that computer.
252b5132
RH
2778
2779A flonum is written by writing (in order)
2780@itemize @bullet
2781@item
2782The digit @samp{0}.
2783@ifset HPPA
2784(@samp{0} is optional on the HPPA.)
2785@end ifset
2786
2787@item
a4fb0134 2788A letter, to tell @command{@value{AS}} the rest of the number is a flonum.
252b5132
RH
2789@ifset GENERIC
2790@kbd{e} is recommended. Case is not important.
2791@ignore
2792@c FIXME: verify if flonum syntax really this vague for most cases
2793(Any otherwise illegal letter works here, but that might be changed. Vax BSD
27944.2 assembler seems to allow any of @samp{defghDEFGH}.)
2795@end ignore
2796
7be1c489 2797On the H8/300, Renesas / SuperH SH,
252b5132
RH
2798and AMD 29K architectures, the letter must be
2799one of the letters @samp{DFPRSX} (in upper or lower case).
2800
2801On the ARC, the letter must be one of the letters @samp{DFRS}
2802(in upper or lower case).
2803
2804On the Intel 960 architecture, the letter must be
2805one of the letters @samp{DFT} (in upper or lower case).
2806
2807On the HPPA architecture, the letter must be @samp{E} (upper case only).
2808@end ifset
2809@ifclear GENERIC
252b5132
RH
2810@ifset ARC
2811One of the letters @samp{DFRS} (in upper or lower case).
2812@end ifset
2813@ifset H8
2814One of the letters @samp{DFPRSX} (in upper or lower case).
2815@end ifset
2816@ifset HPPA
2817The letter @samp{E} (upper case only).
2818@end ifset
2819@ifset I960
2820One of the letters @samp{DFT} (in upper or lower case).
2821@end ifset
2822@end ifclear
2823
2824@item
2825An optional sign: either @samp{+} or @samp{-}.
2826
2827@item
2828An optional @dfn{integer part}: zero or more decimal digits.
2829
2830@item
2831An optional @dfn{fractional part}: @samp{.} followed by zero
2832or more decimal digits.
2833
2834@item
2835An optional exponent, consisting of:
2836
2837@itemize @bullet
2838@item
2839An @samp{E} or @samp{e}.
2840@c I can't find a config where "EXP_CHARS" is other than 'eE', but in
2841@c principle this can perfectly well be different on different targets.
2842@item
2843Optional sign: either @samp{+} or @samp{-}.
2844@item
2845One or more decimal digits.
2846@end itemize
2847
2848@end itemize
2849
2850At least one of the integer part or the fractional part must be
2851present. The floating point number has the usual base-10 value.
2852
a4fb0134 2853@command{@value{AS}} does all processing using integers. Flonums are computed
252b5132 2854independently of any floating point hardware in the computer running
a4fb0134 2855@command{@value{AS}}.
252b5132
RH
2856
2857@ifclear GENERIC
2858@ifset I960
2859@c Bit fields are written as a general facility but are also controlled
2860@c by a conditional-compilation flag---which is as of now (21mar91)
2861@c turned on only by the i960 config of GAS.
2862@node Bit Fields
2863@subsubsection Bit Fields
2864
2865@cindex bit fields
2866@cindex constants, bit field
2867You can also define numeric constants as @dfn{bit fields}.
b45619c0 2868Specify two numbers separated by a colon---
252b5132
RH
2869@example
2870@var{mask}:@var{value}
2871@end example
2872@noindent
a4fb0134 2873@command{@value{AS}} applies a bitwise @sc{and} between @var{mask} and
252b5132
RH
2874@var{value}.
2875
2876The resulting number is then packed
2877@ifset GENERIC
2878@c this conditional paren in case bit fields turned on elsewhere than 960
2879(in host-dependent byte order)
2880@end ifset
2881into a field whose width depends on which assembler directive has the
2882bit-field as its argument. Overflow (a result from the bitwise and
2883requiring more binary digits to represent) is not an error; instead,
2884more constants are generated, of the specified width, beginning with the
2885least significant digits.@refill
2886
2887The directives @code{.byte}, @code{.hword}, @code{.int}, @code{.long},
2888@code{.short}, and @code{.word} accept bit-field arguments.
2889@end ifset
2890@end ifclear
2891
2892@node Sections
2893@chapter Sections and Relocation
2894@cindex sections
2895@cindex relocation
2896
2897@menu
2898* Secs Background:: Background
2899* Ld Sections:: Linker Sections
2900* As Sections:: Assembler Internal Sections
2901* Sub-Sections:: Sub-Sections
2902* bss:: bss Section
2903@end menu
2904
2905@node Secs Background
2906@section Background
2907
2908Roughly, a section is a range of addresses, with no gaps; all data
2909``in'' those addresses is treated the same for some particular purpose.
2910For example there may be a ``read only'' section.
2911
2912@cindex linker, and assembler
2913@cindex assembler, and linker
2914The linker @code{@value{LD}} reads many object files (partial programs) and
a4fb0134 2915combines their contents to form a runnable program. When @command{@value{AS}}
252b5132
RH
2916emits an object file, the partial program is assumed to start at address 0.
2917@code{@value{LD}} assigns the final addresses for the partial program, so that
2918different partial programs do not overlap. This is actually an
a4fb0134 2919oversimplification, but it suffices to explain how @command{@value{AS}} uses
252b5132
RH
2920sections.
2921
2922@code{@value{LD}} moves blocks of bytes of your program to their run-time
2923addresses. These blocks slide to their run-time addresses as rigid
2924units; their length does not change and neither does the order of bytes
2925within them. Such a rigid unit is called a @emph{section}. Assigning
2926run-time addresses to sections is called @dfn{relocation}. It includes
2927the task of adjusting mentions of object-file addresses so they refer to
2928the proper run-time addresses.
2929@ifset H8
7be1c489 2930For the H8/300, and for the Renesas / SuperH SH,
a4fb0134 2931@command{@value{AS}} pads sections if needed to
252b5132
RH
2932ensure they end on a word (sixteen bit) boundary.
2933@end ifset
2934
2935@cindex standard assembler sections
a4fb0134 2936An object file written by @command{@value{AS}} has at least three sections, any
252b5132
RH
2937of which may be empty. These are named @dfn{text}, @dfn{data} and
2938@dfn{bss} sections.
2939
c1253627 2940@ifset COFF-ELF
252b5132 2941@ifset GENERIC
c1253627 2942When it generates COFF or ELF output,
252b5132 2943@end ifset
a4fb0134 2944@command{@value{AS}} can also generate whatever other named sections you specify
252b5132
RH
2945using the @samp{.section} directive (@pxref{Section,,@code{.section}}).
2946If you do not use any directives that place output in the @samp{.text}
2947or @samp{.data} sections, these sections still exist, but are empty.
2948@end ifset
2949
2950@ifset HPPA
2951@ifset GENERIC
a4fb0134 2952When @command{@value{AS}} generates SOM or ELF output for the HPPA,
252b5132 2953@end ifset
a4fb0134 2954@command{@value{AS}} can also generate whatever other named sections you
252b5132
RH
2955specify using the @samp{.space} and @samp{.subspace} directives. See
2956@cite{HP9000 Series 800 Assembly Language Reference Manual}
2957(HP 92432-90001) for details on the @samp{.space} and @samp{.subspace}
2958assembler directives.
2959
2960@ifset SOM
a4fb0134 2961Additionally, @command{@value{AS}} uses different names for the standard
252b5132
RH
2962text, data, and bss sections when generating SOM output. Program text
2963is placed into the @samp{$CODE$} section, data into @samp{$DATA$}, and
2964BSS into @samp{$BSS$}.
2965@end ifset
2966@end ifset
2967
2968Within the object file, the text section starts at address @code{0}, the
2969data section follows, and the bss section follows the data section.
2970
2971@ifset HPPA
2972When generating either SOM or ELF output files on the HPPA, the text
2973section starts at address @code{0}, the data section at address
2974@code{0x4000000}, and the bss section follows the data section.
2975@end ifset
2976
2977To let @code{@value{LD}} know which data changes when the sections are
a4fb0134 2978relocated, and how to change that data, @command{@value{AS}} also writes to the
252b5132
RH
2979object file details of the relocation needed. To perform relocation
2980@code{@value{LD}} must know, each time an address in the object
2981file is mentioned:
2982@itemize @bullet
2983@item
2984Where in the object file is the beginning of this reference to
2985an address?
2986@item
2987How long (in bytes) is this reference?
2988@item
2989Which section does the address refer to? What is the numeric value of
2990@display
2991(@var{address}) @minus{} (@var{start-address of section})?
2992@end display
2993@item
2994Is the reference to an address ``Program-Counter relative''?
2995@end itemize
2996
2997@cindex addresses, format of
2998@cindex section-relative addressing
a4fb0134 2999In fact, every address @command{@value{AS}} ever uses is expressed as
252b5132
RH
3000@display
3001(@var{section}) + (@var{offset into section})
3002@end display
3003@noindent
a4fb0134 3004Further, most expressions @command{@value{AS}} computes have this section-relative
252b5132
RH
3005nature.
3006@ifset SOM
3007(For some object formats, such as SOM for the HPPA, some expressions are
3008symbol-relative instead.)
3009@end ifset
3010
3011In this manual we use the notation @{@var{secname} @var{N}@} to mean ``offset
3012@var{N} into section @var{secname}.''
3013
3014Apart from text, data and bss sections you need to know about the
3015@dfn{absolute} section. When @code{@value{LD}} mixes partial programs,
3016addresses in the absolute section remain unchanged. For example, address
3017@code{@{absolute 0@}} is ``relocated'' to run-time address 0 by
3018@code{@value{LD}}. Although the linker never arranges two partial programs'
3019data sections with overlapping addresses after linking, @emph{by definition}
3020their absolute sections must overlap. Address @code{@{absolute@ 239@}} in one
3021part of a program is always the same address when the program is running as
3022address @code{@{absolute@ 239@}} in any other part of the program.
3023
3024The idea of sections is extended to the @dfn{undefined} section. Any
3025address whose section is unknown at assembly time is by definition
3026rendered @{undefined @var{U}@}---where @var{U} is filled in later.
3027Since numbers are always defined, the only way to generate an undefined
3028address is to mention an undefined symbol. A reference to a named
3029common block would be such a symbol: its value is unknown at assembly
3030time so it has section @emph{undefined}.
3031
3032By analogy the word @emph{section} is used to describe groups of sections in
3033the linked program. @code{@value{LD}} puts all partial programs' text
3034sections in contiguous addresses in the linked program. It is
3035customary to refer to the @emph{text section} of a program, meaning all
3036the addresses of all partial programs' text sections. Likewise for
3037data and bss sections.
3038
3039Some sections are manipulated by @code{@value{LD}}; others are invented for
a4fb0134 3040use of @command{@value{AS}} and have no meaning except during assembly.
252b5132
RH
3041
3042@node Ld Sections
3043@section Linker Sections
3044@code{@value{LD}} deals with just four kinds of sections, summarized below.
3045
3046@table @strong
3047
c1253627 3048@ifset COFF-ELF
252b5132
RH
3049@cindex named sections
3050@cindex sections, named
3051@item named sections
3052@end ifset
3053@ifset aout-bout
3054@cindex text section
3055@cindex data section
3056@itemx text section
3057@itemx data section
3058@end ifset
a4fb0134 3059These sections hold your program. @command{@value{AS}} and @code{@value{LD}} treat them as
252b5132 3060separate but equal sections. Anything you can say of one section is
c1253627
NC
3061true of another.
3062@c @ifset aout-bout
252b5132
RH
3063When the program is running, however, it is
3064customary for the text section to be unalterable. The
3065text section is often shared among processes: it contains
3066instructions, constants and the like. The data section of a running
3067program is usually alterable: for example, C variables would be stored
3068in the data section.
c1253627 3069@c @end ifset
252b5132
RH
3070
3071@cindex bss section
3072@item bss section
3073This section contains zeroed bytes when your program begins running. It
a349d9dd 3074is used to hold uninitialized variables or common storage. The length of
252b5132
RH
3075each partial program's bss section is important, but because it starts
3076out containing zeroed bytes there is no need to store explicit zero
3077bytes in the object file. The bss section was invented to eliminate
3078those explicit zeros from object files.
3079
3080@cindex absolute section
3081@item absolute section
3082Address 0 of this section is always ``relocated'' to runtime address 0.
3083This is useful if you want to refer to an address that @code{@value{LD}} must
3084not change when relocating. In this sense we speak of absolute
3085addresses being ``unrelocatable'': they do not change during relocation.
3086
3087@cindex undefined section
3088@item undefined section
3089This ``section'' is a catch-all for address references to objects not in
3090the preceding sections.
3091@c FIXME: ref to some other doc on obj-file formats could go here.
3092@end table
3093
3094@cindex relocation example
3095An idealized example of three relocatable sections follows.
c1253627 3096@ifset COFF-ELF
252b5132
RH
3097The example uses the traditional section names @samp{.text} and @samp{.data}.
3098@end ifset
3099Memory addresses are on the horizontal axis.
3100
3101@c TEXI2ROFF-KILL
c1253627 3102@ifnottex
252b5132
RH
3103@c END TEXI2ROFF-KILL
3104@smallexample
3105 +-----+----+--+
3106partial program # 1: |ttttt|dddd|00|
3107 +-----+----+--+
3108
3109 text data bss
3110 seg. seg. seg.
3111
3112 +---+---+---+
3113partial program # 2: |TTT|DDD|000|
3114 +---+---+---+
3115
3116 +--+---+-----+--+----+---+-----+~~
3117linked program: | |TTT|ttttt| |dddd|DDD|00000|
3118 +--+---+-----+--+----+---+-----+~~
3119
3120 addresses: 0 @dots{}
3121@end smallexample
3122@c TEXI2ROFF-KILL
c1253627 3123@end ifnottex
252b5132
RH
3124@need 5000
3125@tex
c1253627 3126\bigskip
252b5132
RH
3127\line{\it Partial program \#1: \hfil}
3128\line{\ibox{2.5cm}{\tt text}\ibox{2cm}{\tt data}\ibox{1cm}{\tt bss}\hfil}
3129\line{\boxit{2.5cm}{\tt ttttt}\boxit{2cm}{\tt dddd}\boxit{1cm}{\tt 00}\hfil}
3130
3131\line{\it Partial program \#2: \hfil}
3132\line{\ibox{1cm}{\tt text}\ibox{1.5cm}{\tt data}\ibox{1cm}{\tt bss}\hfil}
3133\line{\boxit{1cm}{\tt TTT}\boxit{1.5cm}{\tt DDDD}\boxit{1cm}{\tt 000}\hfil}
3134
3135\line{\it linked program: \hfil}
3136\line{\ibox{.5cm}{}\ibox{1cm}{\tt text}\ibox{2.5cm}{}\ibox{.75cm}{}\ibox{2cm}{\tt data}\ibox{1.5cm}{}\ibox{2cm}{\tt bss}\hfil}
3137\line{\boxit{.5cm}{}\boxit{1cm}{\tt TTT}\boxit{2.5cm}{\tt
3138ttttt}\boxit{.75cm}{}\boxit{2cm}{\tt dddd}\boxit{1.5cm}{\tt
3139DDDD}\boxit{2cm}{\tt 00000}\ \dots\hfil}
3140
3141\line{\it addresses: \hfil}
3142\line{0\dots\hfil}
3143
3144@end tex
3145@c END TEXI2ROFF-KILL
3146
3147@node As Sections
3148@section Assembler Internal Sections
3149
3150@cindex internal assembler sections
3151@cindex sections in messages, internal
a4fb0134 3152These sections are meant only for the internal use of @command{@value{AS}}. They
252b5132 3153have no meaning at run-time. You do not really need to know about these
a4fb0134 3154sections for most purposes; but they can be mentioned in @command{@value{AS}}
252b5132 3155warning messages, so it might be helpful to have an idea of their
a4fb0134 3156meanings to @command{@value{AS}}. These sections are used to permit the
252b5132
RH
3157value of every expression in your assembly language program to be a
3158section-relative address.
3159
3160@table @b
3161@cindex assembler internal logic error
3162@item ASSEMBLER-INTERNAL-LOGIC-ERROR!
3163An internal assembler logic error has been found. This means there is a
3164bug in the assembler.
3165
3166@cindex expr (internal section)
3167@item expr section
3168The assembler stores complex expression internally as combinations of
3169symbols. When it needs to represent an expression as a symbol, it puts
3170it in the expr section.
3171@c FIXME item debug
3172@c FIXME item transfer[t] vector preload
3173@c FIXME item transfer[t] vector postload
3174@c FIXME item register
3175@end table
3176
3177@node Sub-Sections
3178@section Sub-Sections
3179
3180@cindex numbered subsections
3181@cindex grouping data
3182@ifset aout-bout
3183Assembled bytes
c1253627 3184@ifset COFF-ELF
252b5132
RH
3185conventionally
3186@end ifset
3187fall into two sections: text and data.
3188@end ifset
3189You may have separate groups of
3190@ifset GENERIC
3191data in named sections
3192@end ifset
3193@ifclear GENERIC
3194@ifclear aout-bout
3195data in named sections
3196@end ifclear
3197@ifset aout-bout
3198text or data
3199@end ifset
3200@end ifclear
3201that you want to end up near to each other in the object file, even though they
a4fb0134 3202are not contiguous in the assembler source. @command{@value{AS}} allows you to
252b5132
RH
3203use @dfn{subsections} for this purpose. Within each section, there can be
3204numbered subsections with values from 0 to 8192. Objects assembled into the
3205same subsection go into the object file together with other objects in the same
3206subsection. For example, a compiler might want to store constants in the text
3207section, but might not want to have them interspersed with the program being
3208assembled. In this case, the compiler could issue a @samp{.text 0} before each
3209section of code being output, and a @samp{.text 1} before each group of
3210constants being output.
3211
3212Subsections are optional. If you do not use subsections, everything
3213goes in subsection number zero.
3214
3215@ifset GENERIC
3216Each subsection is zero-padded up to a multiple of four bytes.
3217(Subsections may be padded a different amount on different flavors
a4fb0134 3218of @command{@value{AS}}.)
252b5132
RH
3219@end ifset
3220@ifclear GENERIC
3221@ifset H8
7be1c489 3222On the H8/300 platform, each subsection is zero-padded to a word
252b5132 3223boundary (two bytes).
c2dcd04e 3224The same is true on the Renesas SH.
252b5132
RH
3225@end ifset
3226@ifset I960
3227@c FIXME section padding (alignment)?
3228@c Rich Pixley says padding here depends on target obj code format; that
3229@c doesn't seem particularly useful to say without further elaboration,
3230@c so for now I say nothing about it. If this is a generic BFD issue,
3231@c these paragraphs might need to vanish from this manual, and be
3232@c discussed in BFD chapter of binutils (or some such).
3233@end ifset
252b5132
RH
3234@end ifclear
3235
3236Subsections appear in your object file in numeric order, lowest numbered
3237to highest. (All this to be compatible with other people's assemblers.)
3238The object file contains no representation of subsections; @code{@value{LD}} and
3239other programs that manipulate object files see no trace of them.
3240They just see all your text subsections as a text section, and all your
3241data subsections as a data section.
3242
3243To specify which subsection you want subsequent statements assembled
3244into, use a numeric argument to specify it, in a @samp{.text
3245@var{expression}} or a @samp{.data @var{expression}} statement.
ed9589d4 3246@ifset COFF
252b5132 3247@ifset GENERIC
ed9589d4 3248When generating COFF output, you
252b5132
RH
3249@end ifset
3250@ifclear GENERIC
3251You
3252@end ifclear
3253can also use an extra subsection
3254argument with arbitrary named sections: @samp{.section @var{name},
3255@var{expression}}.
3256@end ifset
ed9589d4
BW
3257@ifset ELF
3258@ifset GENERIC
3259When generating ELF output, you
3260@end ifset
3261@ifclear GENERIC
3262You
3263@end ifclear
3264can also use the @code{.subsection} directive (@pxref{SubSection})
3265to specify a subsection: @samp{.subsection @var{expression}}.
3266@end ifset
96e9638b
BW
3267@var{Expression} should be an absolute expression
3268(@pxref{Expressions}). If you just say @samp{.text} then @samp{.text 0}
252b5132
RH
3269is assumed. Likewise @samp{.data} means @samp{.data 0}. Assembly
3270begins in @code{text 0}. For instance:
3271@smallexample
3272.text 0 # The default subsection is text 0 anyway.
3273.ascii "This lives in the first text subsection. *"
3274.text 1
3275.ascii "But this lives in the second text subsection."
3276.data 0
3277.ascii "This lives in the data section,"
3278.ascii "in the first data subsection."
3279.text 0
3280.ascii "This lives in the first text section,"
3281.ascii "immediately following the asterisk (*)."
3282@end smallexample
3283
3284Each section has a @dfn{location counter} incremented by one for every byte
3285assembled into that section. Because subsections are merely a convenience
a4fb0134 3286restricted to @command{@value{AS}} there is no concept of a subsection location
252b5132
RH
3287counter. There is no way to directly manipulate a location counter---but the
3288@code{.align} directive changes it, and any label definition captures its
3289current value. The location counter of the section where statements are being
3290assembled is said to be the @dfn{active} location counter.
3291
3292@node bss
3293@section bss Section
3294
3295@cindex bss section
3296@cindex common variable storage
3297The bss section is used for local common variable storage.
3298You may allocate address space in the bss section, but you may
3299not dictate data to load into it before your program executes. When
3300your program starts running, all the contents of the bss
3301section are zeroed bytes.
3302
3303The @code{.lcomm} pseudo-op defines a symbol in the bss section; see
3304@ref{Lcomm,,@code{.lcomm}}.
3305
3306The @code{.comm} pseudo-op may be used to declare a common symbol, which is
96e9638b 3307another form of uninitialized symbol; see @ref{Comm,,@code{.comm}}.
252b5132
RH
3308
3309@ifset GENERIC
3310When assembling for a target which supports multiple sections, such as ELF or
3311COFF, you may switch into the @code{.bss} section and define symbols as usual;
3312see @ref{Section,,@code{.section}}. You may only assemble zero values into the
3313section. Typically the section will only contain symbol definitions and
3314@code{.skip} directives (@pxref{Skip,,@code{.skip}}).
3315@end ifset
3316
3317@node Symbols
3318@chapter Symbols
3319
3320@cindex symbols
3321Symbols are a central concept: the programmer uses symbols to name
3322things, the linker uses symbols to link, and the debugger uses symbols
3323to debug.
3324
3325@quotation
3326@cindex debuggers, and symbol order
a4fb0134 3327@emph{Warning:} @command{@value{AS}} does not place symbols in the object file in
252b5132
RH
3328the same order they were declared. This may break some debuggers.
3329@end quotation
3330
3331@menu
3332* Labels:: Labels
3333* Setting Symbols:: Giving Symbols Other Values
3334* Symbol Names:: Symbol Names
3335* Dot:: The Special Dot Symbol
3336* Symbol Attributes:: Symbol Attributes
3337@end menu
3338
3339@node Labels
3340@section Labels
3341
3342@cindex labels
3343A @dfn{label} is written as a symbol immediately followed by a colon
3344@samp{:}. The symbol then represents the current value of the
3345active location counter, and is, for example, a suitable instruction
3346operand. You are warned if you use the same symbol to represent two
3347different locations: the first definition overrides any other
3348definitions.
3349
3350@ifset HPPA
3351On the HPPA, the usual form for a label need not be immediately followed by a
3352colon, but instead must start in column zero. Only one label may be defined on
a4fb0134 3353a single line. To work around this, the HPPA version of @command{@value{AS}} also
252b5132
RH
3354provides a special directive @code{.label} for defining labels more flexibly.
3355@end ifset
3356
3357@node Setting Symbols
3358@section Giving Symbols Other Values
3359
3360@cindex assigning values to symbols
3361@cindex symbol values, assigning
3362A symbol can be given an arbitrary value by writing a symbol, followed
3363by an equals sign @samp{=}, followed by an expression
3364(@pxref{Expressions}). This is equivalent to using the @code{.set}
9497f5ac
NC
3365directive. @xref{Set,,@code{.set}}. In the same way, using a double
3366equals sign @samp{=}@samp{=} here represents an equivalent of the
3367@code{.eqv} directive. @xref{Eqv,,@code{.eqv}}.
252b5132 3368
f8739b83
JZ
3369@ifset Blackfin
3370Blackfin does not support symbol assignment with @samp{=}.
3371@end ifset
3372
252b5132
RH
3373@node Symbol Names
3374@section Symbol Names
3375
3376@cindex symbol names
3377@cindex names, symbol
3378@ifclear SPECIAL-SYMS
3379Symbol names begin with a letter or with one of @samp{._}. On most
3380machines, you can also use @code{$} in symbol names; exceptions are
3381noted in @ref{Machine Dependencies}. That character may be followed by any
96e9638b
BW
3382string of digits, letters, dollar signs (unless otherwise noted for a
3383particular target machine), and underscores.
252b5132 3384@end ifclear
252b5132
RH
3385@ifset SPECIAL-SYMS
3386@ifset H8
3387Symbol names begin with a letter or with one of @samp{._}. On the
7be1c489 3388Renesas SH you can also use @code{$} in symbol names. That
c2dcd04e
NC
3389character may be followed by any string of digits, letters, dollar signs (save
3390on the H8/300), and underscores.
252b5132
RH
3391@end ifset
3392@end ifset
3393
3394Case of letters is significant: @code{foo} is a different symbol name
3395than @code{Foo}.
3396
3397Each symbol has exactly one name. Each name in an assembly language program
3398refers to exactly one symbol. You may use that symbol name any number of times
3399in a program.
3400
3401@subheading Local Symbol Names
3402
3403@cindex local symbol names
3404@cindex symbol names, local
ba83aca1
BW
3405A local symbol is any symbol beginning with certain local label prefixes.
3406By default, the local label prefix is @samp{.L} for ELF systems or
3407@samp{L} for traditional a.out systems, but each target may have its own
3408set of local label prefixes.
3409@ifset HPPA
3410On the HPPA local symbols begin with @samp{L$}.
3411@end ifset
3412
3413Local symbols are defined and used within the assembler, but they are
3414normally not saved in object files. Thus, they are not visible when debugging.
3415You may use the @samp{-L} option (@pxref{L, ,Include Local Symbols:
3416@option{-L}}) to retain the local symbols in the object files.
3417
3418@subheading Local Labels
3419
3420@cindex local labels
252b5132
RH
3421@cindex temporary symbol names
3422@cindex symbol names, temporary
ba83aca1 3423Local labels help compilers and programmers use names temporarily.
2d5aaba0
NC
3424They create symbols which are guaranteed to be unique over the entire scope of
3425the input source code and which can be referred to by a simple notation.
ba83aca1 3426To define a local label, write a label of the form @samp{@b{N}:} (where @b{N}
2d5aaba0 3427represents any positive integer). To refer to the most recent previous
ba83aca1 3428definition of that label write @samp{@b{N}b}, using the same number as when
2d5aaba0 3429you defined the label. To refer to the next definition of a local label, write
96e9638b 3430@samp{@b{N}f}---the @samp{b} stands for ``backwards'' and the @samp{f} stands
2d5aaba0
NC
3431for ``forwards''.
3432
3433There is no restriction on how you can use these labels, and you can reuse them
3434too. So that it is possible to repeatedly define the same local label (using
3435the same number @samp{@b{N}}), although you can only refer to the most recently
3436defined local label of that number (for a backwards reference) or the next
3437definition of a specific local label for a forward reference. It is also worth
3438noting that the first 10 local labels (@samp{@b{0:}}@dots{}@samp{@b{9:}}) are
3439implemented in a slightly more efficient manner than the others.
3440
3441Here is an example:
3442
3443@smallexample
34441: branch 1f
34452: branch 1b
34461: branch 2f
34472: branch 1b
3448@end smallexample
3449
3450Which is the equivalent of:
3451
3452@smallexample
3453label_1: branch label_3
3454label_2: branch label_1
3455label_3: branch label_4
3456label_4: branch label_3
3457@end smallexample
3458
ba83aca1 3459Local label names are only a notational device. They are immediately
2d5aaba0 3460transformed into more conventional symbol names before the assembler uses them.
96e9638b
BW
3461The symbol names are stored in the symbol table, appear in error messages, and
3462are optionally emitted to the object file. The names are constructed using
3463these parts:
252b5132
RH
3464
3465@table @code
ba83aca1
BW
3466@item @emph{local label prefix}
3467All local symbols begin with the system-specific local label prefix.
3468Normally both @command{@value{AS}} and @code{@value{LD}} forget symbols
3469that start with the local label prefix. These labels are
252b5132 3470used for symbols you are never intended to see. If you use the
a4fb0134 3471@samp{-L} option then @command{@value{AS}} retains these symbols in the
252b5132
RH
3472object file. If you also instruct @code{@value{LD}} to retain these symbols,
3473you may use them in debugging.
3474
2d5aaba0
NC
3475@item @var{number}
3476This is the number that was used in the local label definition. So if the
01642c12 3477label is written @samp{55:} then the number is @samp{55}.
252b5132 3478
2d5aaba0
NC
3479@item @kbd{C-B}
3480This unusual character is included so you do not accidentally invent a symbol
3481of the same name. The character has ASCII value of @samp{\002} (control-B).
252b5132
RH
3482
3483@item @emph{ordinal number}
2d5aaba0 3484This is a serial number to keep the labels distinct. The first definition of
01642c12 3485@samp{0:} gets the number @samp{1}. The 15th definition of @samp{0:} gets the
2d5aaba0 3486number @samp{15}, and so on. Likewise the first definition of @samp{1:} gets
b45619c0 3487the number @samp{1} and its 15th definition gets @samp{15} as well.
252b5132
RH
3488@end table
3489
ba83aca1
BW
3490So for example, the first @code{1:} may be named @code{.L1@kbd{C-B}1}, and
3491the 44th @code{3:} may be named @code{.L3@kbd{C-B}44}.
2d5aaba0
NC
3492
3493@subheading Dollar Local Labels
3494@cindex dollar local symbols
3495
3496@code{@value{AS}} also supports an even more local form of local labels called
96e9638b
BW
3497dollar labels. These labels go out of scope (i.e., they become undefined) as
3498soon as a non-local label is defined. Thus they remain valid for only a small
2d5aaba0
NC
3499region of the input source code. Normal local labels, by contrast, remain in
3500scope for the entire file, or until they are redefined by another occurrence of
3501the same local label.
3502
3503Dollar labels are defined in exactly the same way as ordinary local labels,
77cca80f
NC
3504except that they have a dollar sign suffix to their numeric value, e.g.,
3505@samp{@b{55$:}}.
2d5aaba0
NC
3506
3507They can also be distinguished from ordinary local labels by their transformed
96e9638b
BW
3508names which use ASCII character @samp{\001} (control-A) as the magic character
3509to distinguish them from ordinary labels. For example, the fifth definition of
ba83aca1 3510@samp{6$} may be named @samp{.L6@kbd{C-A}5}.
252b5132
RH
3511
3512@node Dot
3513@section The Special Dot Symbol
3514
3515@cindex dot (symbol)
3516@cindex @code{.} (symbol)
3517@cindex current address
3518@cindex location counter
3519The special symbol @samp{.} refers to the current address that
a4fb0134 3520@command{@value{AS}} is assembling into. Thus, the expression @samp{melvin:
252b5132
RH
3521.long .} defines @code{melvin} to contain its own address.
3522Assigning a value to @code{.} is treated the same as a @code{.org}
884f0d36 3523directive.
252b5132 3524@ifclear no-space-dir
884f0d36 3525Thus, the expression @samp{.=.+4} is the same as saying
252b5132
RH
3526@samp{.space 4}.
3527@end ifclear
252b5132
RH
3528
3529@node Symbol Attributes
3530@section Symbol Attributes
3531
3532@cindex symbol attributes
3533@cindex attributes, symbol
3534Every symbol has, as well as its name, the attributes ``Value'' and
3535``Type''. Depending on output format, symbols can also have auxiliary
3536attributes.
3537@ifset INTERNALS
3538The detailed definitions are in @file{a.out.h}.
3539@end ifset
3540
a4fb0134 3541If you use a symbol without defining it, @command{@value{AS}} assumes zero for
252b5132
RH
3542all these attributes, and probably won't warn you. This makes the
3543symbol an externally defined symbol, which is generally what you
3544would want.
3545
3546@menu
3547* Symbol Value:: Value
3548* Symbol Type:: Type
3549@ifset aout-bout
3550@ifset GENERIC
3551* a.out Symbols:: Symbol Attributes: @code{a.out}
3552@end ifset
3553@ifclear GENERIC
3554@ifclear BOUT
3555* a.out Symbols:: Symbol Attributes: @code{a.out}
3556@end ifclear
3557@ifset BOUT
3558* a.out Symbols:: Symbol Attributes: @code{a.out}, @code{b.out}
3559@end ifset
3560@end ifclear
3561@end ifset
3562@ifset COFF
3563* COFF Symbols:: Symbol Attributes for COFF
3564@end ifset
3565@ifset SOM
3566* SOM Symbols:: Symbol Attributes for SOM
3567@end ifset
3568@end menu
3569
3570@node Symbol Value
3571@subsection Value
3572
3573@cindex value of a symbol
3574@cindex symbol value
3575The value of a symbol is (usually) 32 bits. For a symbol which labels a
3576location in the text, data, bss or absolute sections the value is the
3577number of addresses from the start of that section to the label.
3578Naturally for text, data and bss sections the value of a symbol changes
3579as @code{@value{LD}} changes section base addresses during linking. Absolute
3580symbols' values do not change during linking: that is why they are
3581called absolute.
3582
3583The value of an undefined symbol is treated in a special way. If it is
35840 then the symbol is not defined in this assembler source file, and
3585@code{@value{LD}} tries to determine its value from other files linked into the
3586same program. You make this kind of symbol simply by mentioning a symbol
3587name without defining it. A non-zero value represents a @code{.comm}
3588common declaration. The value is how much common storage to reserve, in
3589bytes (addresses). The symbol refers to the first address of the
3590allocated storage.
3591
3592@node Symbol Type
3593@subsection Type
3594
3595@cindex type of a symbol
3596@cindex symbol type
3597The type attribute of a symbol contains relocation (section)
3598information, any flag settings indicating that a symbol is external, and
3599(optionally), other information for linkers and debuggers. The exact
3600format depends on the object-code output format in use.
3601
3602@ifset aout-bout
3603@ifclear GENERIC
3604@ifset BOUT
3605@c The following avoids a "widow" subsection title. @group would be
3606@c better if it were available outside examples.
3607@need 1000
3608@node a.out Symbols
3609@subsection Symbol Attributes: @code{a.out}, @code{b.out}
3610
3611@cindex @code{b.out} symbol attributes
3612@cindex symbol attributes, @code{b.out}
a4fb0134 3613These symbol attributes appear only when @command{@value{AS}} is configured for
252b5132
RH
3614one of the Berkeley-descended object output formats---@code{a.out} or
3615@code{b.out}.
3616
3617@end ifset
3618@ifclear BOUT
3619@node a.out Symbols
3620@subsection Symbol Attributes: @code{a.out}
3621
3622@cindex @code{a.out} symbol attributes
3623@cindex symbol attributes, @code{a.out}
3624
3625@end ifclear
3626@end ifclear
3627@ifset GENERIC
3628@node a.out Symbols
3629@subsection Symbol Attributes: @code{a.out}
3630
3631@cindex @code{a.out} symbol attributes
3632@cindex symbol attributes, @code{a.out}
3633
3634@end ifset
3635@menu
3636* Symbol Desc:: Descriptor
3637* Symbol Other:: Other
3638@end menu
3639
3640@node Symbol Desc
3641@subsubsection Descriptor
3642
3643@cindex descriptor, of @code{a.out} symbol
3644This is an arbitrary 16-bit value. You may establish a symbol's
3645descriptor value by using a @code{.desc} statement
3646(@pxref{Desc,,@code{.desc}}). A descriptor value means nothing to
a4fb0134 3647@command{@value{AS}}.
252b5132
RH
3648
3649@node Symbol Other
3650@subsubsection Other
3651
3652@cindex other attribute, of @code{a.out} symbol
a4fb0134 3653This is an arbitrary 8-bit value. It means nothing to @command{@value{AS}}.
252b5132
RH
3654@end ifset
3655
3656@ifset COFF
3657@node COFF Symbols
3658@subsection Symbol Attributes for COFF
3659
3660@cindex COFF symbol attributes
3661@cindex symbol attributes, COFF
3662
3663The COFF format supports a multitude of auxiliary symbol attributes;
3664like the primary symbol attributes, they are set between @code{.def} and
3665@code{.endef} directives.
3666
3667@subsubsection Primary Attributes
3668
3669@cindex primary attributes, COFF symbols
3670The symbol name is set with @code{.def}; the value and type,
3671respectively, with @code{.val} and @code{.type}.
3672
3673@subsubsection Auxiliary Attributes
3674
3675@cindex auxiliary attributes, COFF symbols
a4fb0134 3676The @command{@value{AS}} directives @code{.dim}, @code{.line}, @code{.scl},
c87db184
CF
3677@code{.size}, @code{.tag}, and @code{.weak} can generate auxiliary symbol
3678table information for COFF.
252b5132
RH
3679@end ifset
3680
3681@ifset SOM
3682@node SOM Symbols
3683@subsection Symbol Attributes for SOM
3684
3685@cindex SOM symbol attributes
3686@cindex symbol attributes, SOM
3687
3688The SOM format for the HPPA supports a multitude of symbol attributes set with
3689the @code{.EXPORT} and @code{.IMPORT} directives.
3690
01642c12 3691The attributes are described in @cite{HP9000 Series 800 Assembly
252b5132
RH
3692Language Reference Manual} (HP 92432-90001) under the @code{IMPORT} and
3693@code{EXPORT} assembler directive documentation.
3694@end ifset
3695
3696@node Expressions
3697@chapter Expressions
3698
3699@cindex expressions
3700@cindex addresses
3701@cindex numeric values
3702An @dfn{expression} specifies an address or numeric value.
3703Whitespace may precede and/or follow an expression.
3704
3705The result of an expression must be an absolute number, or else an offset into
3706a particular section. If an expression is not absolute, and there is not
a4fb0134 3707enough information when @command{@value{AS}} sees the expression to know its
252b5132
RH
3708section, a second pass over the source program might be necessary to interpret
3709the expression---but the second pass is currently not implemented.
a4fb0134 3710@command{@value{AS}} aborts with an error message in this situation.
252b5132
RH
3711
3712@menu
3713* Empty Exprs:: Empty Expressions
3714* Integer Exprs:: Integer Expressions
3715@end menu
3716
3717@node Empty Exprs
3718@section Empty Expressions
3719
3720@cindex empty expressions
3721@cindex expressions, empty
3722An empty expression has no value: it is just whitespace or null.
3723Wherever an absolute expression is required, you may omit the
a4fb0134 3724expression, and @command{@value{AS}} assumes a value of (absolute) 0. This
252b5132
RH
3725is compatible with other assemblers.
3726
3727@node Integer Exprs
3728@section Integer Expressions
3729
3730@cindex integer expressions
3731@cindex expressions, integer
3732An @dfn{integer expression} is one or more @emph{arguments} delimited
3733by @emph{operators}.
3734
3735@menu
3736* Arguments:: Arguments
3737* Operators:: Operators
3738* Prefix Ops:: Prefix Operators
3739* Infix Ops:: Infix Operators
3740@end menu
3741
3742@node Arguments
3743@subsection Arguments
3744
3745@cindex expression arguments
3746@cindex arguments in expressions
3747@cindex operands in expressions
3748@cindex arithmetic operands
3749@dfn{Arguments} are symbols, numbers or subexpressions. In other
3750contexts arguments are sometimes called ``arithmetic operands''. In
3751this manual, to avoid confusing them with the ``instruction operands'' of
3752the machine language, we use the term ``argument'' to refer to parts of
3753expressions only, reserving the word ``operand'' to refer only to machine
3754instruction operands.
3755
3756Symbols are evaluated to yield @{@var{section} @var{NNN}@} where
3757@var{section} is one of text, data, bss, absolute,
3758or undefined. @var{NNN} is a signed, 2's complement 32 bit
3759integer.
3760
3761Numbers are usually integers.
3762
3763A number can be a flonum or bignum. In this case, you are warned
a4fb0134 3764that only the low order 32 bits are used, and @command{@value{AS}} pretends
252b5132
RH
3765these 32 bits are an integer. You may write integer-manipulating
3766instructions that act on exotic constants, compatible with other
3767assemblers.
3768
3769@cindex subexpressions
3770Subexpressions are a left parenthesis @samp{(} followed by an integer
3771expression, followed by a right parenthesis @samp{)}; or a prefix
3772operator followed by an argument.
3773
3774@node Operators
3775@subsection Operators
3776
3777@cindex operators, in expressions
3778@cindex arithmetic functions
3779@cindex functions, in expressions
3780@dfn{Operators} are arithmetic functions, like @code{+} or @code{%}. Prefix
3781operators are followed by an argument. Infix operators appear
3782between their arguments. Operators may be preceded and/or followed by
3783whitespace.
3784
3785@node Prefix Ops
3786@subsection Prefix Operator
3787
3788@cindex prefix operators
a4fb0134 3789@command{@value{AS}} has the following @dfn{prefix operators}. They each take
252b5132
RH
3790one argument, which must be absolute.
3791
3792@c the tex/end tex stuff surrounding this small table is meant to make
3793@c it align, on the printed page, with the similar table in the next
3794@c section (which is inside an enumerate).
3795@tex
3796\global\advance\leftskip by \itemindent
3797@end tex
3798
3799@table @code
3800@item -
3801@dfn{Negation}. Two's complement negation.
3802@item ~
3803@dfn{Complementation}. Bitwise not.
3804@end table
3805
3806@tex
3807\global\advance\leftskip by -\itemindent
3808@end tex
3809
3810@node Infix Ops
3811@subsection Infix Operators
3812
3813@cindex infix operators
3814@cindex operators, permitted arguments
3815@dfn{Infix operators} take two arguments, one on either side. Operators
3816have precedence, but operations with equal precedence are performed left
a4fb0134 3817to right. Apart from @code{+} or @option{-}, both arguments must be
252b5132
RH
3818absolute, and the result is absolute.
3819
3820@enumerate
3821@cindex operator precedence
3822@cindex precedence of operators
3823
3824@item
3825Highest Precedence
3826
3827@table @code
3828@item *
3829@dfn{Multiplication}.
3830
3831@item /
3832@dfn{Division}. Truncation is the same as the C operator @samp{/}
3833
3834@item %
3835@dfn{Remainder}.
3836
d1eac9d9 3837@item <<
252b5132
RH
3838@dfn{Shift Left}. Same as the C operator @samp{<<}.
3839
d1eac9d9 3840@item >>
252b5132
RH
3841@dfn{Shift Right}. Same as the C operator @samp{>>}.
3842@end table
3843
3844@item
3845Intermediate precedence
3846
3847@table @code
3848@item |
3849
3850@dfn{Bitwise Inclusive Or}.
3851
3852@item &
3853@dfn{Bitwise And}.
3854
3855@item ^
3856@dfn{Bitwise Exclusive Or}.
3857
3858@item !
3859@dfn{Bitwise Or Not}.
3860@end table
3861
3862@item
b131d4dc 3863Low Precedence
252b5132
RH
3864
3865@table @code
3866@cindex addition, permitted arguments
3867@cindex plus, permitted arguments
3868@cindex arguments for addition
3869@item +
3870@dfn{Addition}. If either argument is absolute, the result has the section of
3871the other argument. You may not add together arguments from different
3872sections.
3873
3874@cindex subtraction, permitted arguments
3875@cindex minus, permitted arguments
3876@cindex arguments for subtraction
3877@item -
3878@dfn{Subtraction}. If the right argument is absolute, the
3879result has the section of the left argument.
3880If both arguments are in the same section, the result is absolute.
3881You may not subtract arguments from different sections.
3882@c FIXME is there still something useful to say about undefined - undefined ?
b131d4dc
NC
3883
3884@cindex comparison expressions
3885@cindex expressions, comparison
3886@item ==
3887@dfn{Is Equal To}
3888@item <>
723a8472 3889@itemx !=
b131d4dc
NC
3890@dfn{Is Not Equal To}
3891@item <
3892@dfn{Is Less Than}
d1eac9d9 3893@item >
b131d4dc 3894@dfn{Is Greater Than}
d1eac9d9 3895@item >=
b131d4dc 3896@dfn{Is Greater Than Or Equal To}
d1eac9d9 3897@item <=
b131d4dc
NC
3898@dfn{Is Less Than Or Equal To}
3899
3900The comparison operators can be used as infix operators. A true results has a
3901value of -1 whereas a false result has a value of 0. Note, these operators
3902perform signed comparisons.
3903@end table
3904
3905@item Lowest Precedence
3906
3907@table @code
3908@item &&
3909@dfn{Logical And}.
3910
3911@item ||
3912@dfn{Logical Or}.
3913
3914These two logical operations can be used to combine the results of sub
3915expressions. Note, unlike the comparison operators a true result returns a
3916value of 1 but a false results does still return 0. Also note that the logical
3917or operator has a slightly lower precedence than logical and.
3918
252b5132
RH
3919@end table
3920@end enumerate
3921
3922In short, it's only meaningful to add or subtract the @emph{offsets} in an
3923address; you can only have a defined section in one of the two arguments.
3924
3925@node Pseudo Ops
3926@chapter Assembler Directives
3927
3928@cindex directives, machine independent
3929@cindex pseudo-ops, machine independent
3930@cindex machine independent directives
3931All assembler directives have names that begin with a period (@samp{.}).
3932The rest of the name is letters, usually in lower case.
3933
3934This chapter discusses directives that are available regardless of the
3935target machine configuration for the @sc{gnu} assembler.
3936@ifset GENERIC
3937Some machine configurations provide additional directives.
3938@xref{Machine Dependencies}.
3939@end ifset
3940@ifclear GENERIC
3941@ifset machine-directives
96e9638b 3942@xref{Machine Dependencies}, for additional directives.
252b5132
RH
3943@end ifset
3944@end ifclear
3945
3946@menu
3947* Abort:: @code{.abort}
3948@ifset COFF
38a57ae7 3949* ABORT (COFF):: @code{.ABORT}
252b5132 3950@end ifset
f0dc282c 3951
252b5132 3952* Align:: @code{.align @var{abs-expr} , @var{abs-expr}}
caa32fe5 3953* Altmacro:: @code{.altmacro}
252b5132
RH
3954* Ascii:: @code{.ascii "@var{string}"}@dots{}
3955* Asciz:: @code{.asciz "@var{string}"}@dots{}
3956* Balign:: @code{.balign @var{abs-expr} , @var{abs-expr}}
3957* Byte:: @code{.byte @var{expressions}}
4b7d318b 3958* CFI directives:: @code{.cfi_startproc [simple]}, @code{.cfi_endproc}, etc.
ccf8a69b 3959* Comm:: @code{.comm @var{symbol} , @var{length} }
252b5132
RH
3960* Data:: @code{.data @var{subsection}}
3961@ifset COFF
3962* Def:: @code{.def @var{name}}
3963@end ifset
3964@ifset aout-bout
3965* Desc:: @code{.desc @var{symbol}, @var{abs-expression}}
3966@end ifset
3967@ifset COFF
3968* Dim:: @code{.dim}
3969@end ifset
f0dc282c 3970
252b5132
RH
3971* Double:: @code{.double @var{flonums}}
3972* Eject:: @code{.eject}
3973* Else:: @code{.else}
3fd9f047 3974* Elseif:: @code{.elseif}
252b5132
RH
3975* End:: @code{.end}
3976@ifset COFF
3977* Endef:: @code{.endef}
3978@end ifset
f0dc282c 3979
252b5132
RH
3980* Endfunc:: @code{.endfunc}
3981* Endif:: @code{.endif}
3982* Equ:: @code{.equ @var{symbol}, @var{expression}}
3983* Equiv:: @code{.equiv @var{symbol}, @var{expression}}
9497f5ac 3984* Eqv:: @code{.eqv @var{symbol}, @var{expression}}
252b5132 3985* Err:: @code{.err}
d190d046 3986* Error:: @code{.error @var{string}}
252b5132
RH
3987* Exitm:: @code{.exitm}
3988* Extern:: @code{.extern}
3989* Fail:: @code{.fail}
14082c76 3990* File:: @code{.file}
252b5132
RH
3991* Fill:: @code{.fill @var{repeat} , @var{size} , @var{value}}
3992* Float:: @code{.float @var{flonums}}
01642c12 3993* Func:: @code{.func}
252b5132 3994* Global:: @code{.global @var{symbol}}, @code{.globl @var{symbol}}
c91d2e08 3995@ifset ELF
3a99f02f 3996* Gnu_attribute:: @code{.gnu_attribute @var{tag},@var{value}}
c91d2e08
NC
3997* Hidden:: @code{.hidden @var{names}}
3998@end ifset
f0dc282c 3999
252b5132
RH
4000* hword:: @code{.hword @var{expressions}}
4001* Ident:: @code{.ident}
4002* If:: @code{.if @var{absolute expression}}
7e005732 4003* Incbin:: @code{.incbin "@var{file}"[,@var{skip}[,@var{count}]]}
252b5132
RH
4004* Include:: @code{.include "@var{file}"}
4005* Int:: @code{.int @var{expressions}}
c91d2e08
NC
4006@ifset ELF
4007* Internal:: @code{.internal @var{names}}
4008@end ifset
f0dc282c 4009
252b5132
RH
4010* Irp:: @code{.irp @var{symbol},@var{values}}@dots{}
4011* Irpc:: @code{.irpc @var{symbol},@var{values}}@dots{}
4012* Lcomm:: @code{.lcomm @var{symbol} , @var{length}}
4013* Lflags:: @code{.lflags}
4014@ifclear no-line-dir
4015* Line:: @code{.line @var{line-number}}
4016@end ifclear
f0dc282c 4017
252b5132
RH
4018* Linkonce:: @code{.linkonce [@var{type}]}
4019* List:: @code{.list}
bd0eb99b 4020* Ln:: @code{.ln @var{line-number}}
14082c76
BW
4021* Loc:: @code{.loc @var{fileno} @var{lineno}}
4022* Loc_mark_labels:: @code{.loc_mark_labels @var{enable}}
4d4175af
BW
4023@ifset ELF
4024* Local:: @code{.local @var{names}}
4025@end ifset
bd0eb99b 4026
252b5132
RH
4027* Long:: @code{.long @var{expressions}}
4028@ignore
4029* Lsym:: @code{.lsym @var{symbol}, @var{expression}}
4030@end ignore
f0dc282c 4031
252b5132
RH
4032* Macro:: @code{.macro @var{name} @var{args}}@dots{}
4033* MRI:: @code{.mri @var{val}}
caa32fe5 4034* Noaltmacro:: @code{.noaltmacro}
252b5132
RH
4035* Nolist:: @code{.nolist}
4036* Octa:: @code{.octa @var{bignums}}
85234291
L
4037* Org:: @code{.org @var{new-lc}, @var{fill}}
4038* P2align:: @code{.p2align @var{abs-expr}, @var{abs-expr}, @var{abs-expr}}
c91d2e08
NC
4039@ifset ELF
4040* PopSection:: @code{.popsection}
4041* Previous:: @code{.previous}
4042@end ifset
f0dc282c 4043
252b5132 4044* Print:: @code{.print @var{string}}
c91d2e08
NC
4045@ifset ELF
4046* Protected:: @code{.protected @var{names}}
4047@end ifset
f0dc282c 4048
252b5132
RH
4049* Psize:: @code{.psize @var{lines}, @var{columns}}
4050* Purgem:: @code{.purgem @var{name}}
c91d2e08
NC
4051@ifset ELF
4052* PushSection:: @code{.pushsection @var{name}}
4053@end ifset
f0dc282c 4054
252b5132 4055* Quad:: @code{.quad @var{bignums}}
05e9452c 4056* Reloc:: @code{.reloc @var{offset}, @var{reloc_name}[, @var{expression}]}
252b5132
RH
4057* Rept:: @code{.rept @var{count}}
4058* Sbttl:: @code{.sbttl "@var{subheading}"}
4059@ifset COFF
4060* Scl:: @code{.scl @var{class}}
c1253627
NC
4061@end ifset
4062@ifset COFF-ELF
7337fc21 4063* Section:: @code{.section @var{name}[, @var{flags}]}
252b5132 4064@end ifset
f0dc282c 4065
252b5132
RH
4066* Set:: @code{.set @var{symbol}, @var{expression}}
4067* Short:: @code{.short @var{expressions}}
4068* Single:: @code{.single @var{flonums}}
c1253627 4069@ifset COFF-ELF
c91d2e08 4070* Size:: @code{.size [@var{name} , @var{expression}]}
c1253627 4071@end ifset
884f0d36 4072@ifclear no-space-dir
252b5132 4073* Skip:: @code{.skip @var{size} , @var{fill}}
884f0d36
BW
4074@end ifclear
4075
252b5132 4076* Sleb128:: @code{.sleb128 @var{expressions}}
884f0d36 4077@ifclear no-space-dir
252b5132 4078* Space:: @code{.space @var{size} , @var{fill}}
884f0d36 4079@end ifclear
252b5132
RH
4080@ifset have-stabs
4081* Stab:: @code{.stabd, .stabn, .stabs}
4082@end ifset
f0dc282c 4083
38a57ae7 4084* String:: @code{.string "@var{str}"}, @code{.string8 "@var{str}"}, @code{.string16 "@var{str}"}, @code{.string32 "@var{str}"}, @code{.string64 "@var{str}"}
252b5132
RH
4085* Struct:: @code{.struct @var{expression}}
4086@ifset ELF
c91d2e08 4087* SubSection:: @code{.subsection}
252b5132
RH
4088* Symver:: @code{.symver @var{name},@var{name2@@nodename}}
4089@end ifset
f0dc282c 4090
252b5132
RH
4091@ifset COFF
4092* Tag:: @code{.tag @var{structname}}
4093@end ifset
f0dc282c 4094
252b5132
RH
4095* Text:: @code{.text @var{subsection}}
4096* Title:: @code{.title "@var{heading}"}
c1253627 4097@ifset COFF-ELF
c91d2e08 4098* Type:: @code{.type <@var{int} | @var{name} , @var{type description}>}
c1253627
NC
4099@end ifset
4100
c91d2e08 4101* Uleb128:: @code{.uleb128 @var{expressions}}
252b5132 4102@ifset COFF
252b5132
RH
4103* Val:: @code{.val @var{addr}}
4104@end ifset
f0dc282c 4105
2e13b764 4106@ifset ELF
c91d2e08 4107* Version:: @code{.version "@var{string}"}
c91d2e08
NC
4108* VTableEntry:: @code{.vtable_entry @var{table}, @var{offset}}
4109* VTableInherit:: @code{.vtable_inherit @var{child}, @var{parent}}
2e13b764 4110@end ifset
f0dc282c 4111
d190d046 4112* Warning:: @code{.warning @var{string}}
c87db184 4113* Weak:: @code{.weak @var{names}}
06e77878 4114* Weakref:: @code{.weakref @var{alias}, @var{symbol}}
252b5132
RH
4115* Word:: @code{.word @var{expressions}}
4116* Deprecated:: Deprecated Directives
4117@end menu
4118
4119@node Abort
4120@section @code{.abort}
4121
4122@cindex @code{abort} directive
4123@cindex stopping the assembly
4124This directive stops the assembly immediately. It is for
4125compatibility with other assemblers. The original idea was that the
4126assembly language source would be piped into the assembler. If the sender
a4fb0134 4127of the source quit, it could use this directive tells @command{@value{AS}} to
252b5132
RH
4128quit also. One day @code{.abort} will not be supported.
4129
4130@ifset COFF
370b66a1
CD
4131@node ABORT (COFF)
4132@section @code{.ABORT} (COFF)
252b5132
RH
4133
4134@cindex @code{ABORT} directive
a4fb0134 4135When producing COFF output, @command{@value{AS}} accepts this directive as a
252b5132
RH
4136synonym for @samp{.abort}.
4137
4138@ifset BOUT
a4fb0134 4139When producing @code{b.out} output, @command{@value{AS}} accepts this directive,
252b5132
RH
4140but ignores it.
4141@end ifset
4142@end ifset
4143
4144@node Align
4145@section @code{.align @var{abs-expr}, @var{abs-expr}, @var{abs-expr}}
4146
4147@cindex padding the location counter
4148@cindex @code{align} directive
4149Pad the location counter (in the current subsection) to a particular storage
4150boundary. The first expression (which must be absolute) is the alignment
4151required, as described below.
4152
4153The second expression (also absolute) gives the fill value to be stored in the
4154padding bytes. It (and the comma) may be omitted. If it is omitted, the
4155padding bytes are normally zero. However, on some systems, if the section is
4156marked as containing code and the fill value is omitted, the space is filled
4157with no-op instructions.
4158
4159The third expression is also absolute, and is also optional. If it is present,
4160it is the maximum number of bytes that should be skipped by this alignment
4161directive. If doing the alignment would require skipping more bytes than the
4162specified maximum, then the alignment is not done at all. You can omit the
4163fill value (the second argument) entirely by simply using two commas after the
4164required alignment; this can be useful if you want the alignment to be filled
4165with no-op instructions when appropriate.
4166
4167The way the required alignment is specified varies from system to system.
7be1c489 4168For the arc, hppa, i386 using ELF, i860, iq2000, m68k, or32,
60946ad0 4169s390, sparc, tic4x, tic80 and xtensa, the first expression is the
252b5132
RH
4170alignment request in bytes. For example @samp{.align 8} advances
4171the location counter until it is a multiple of 8. If the location counter
60946ad0
AM
4172is already a multiple of 8, no change is needed. For the tic54x, the
4173first expression is the alignment request in words.
252b5132 4174
9e9a9798 4175For other systems, including ppc, i386 using a.out format, arm and
adcf07e6 4176strongarm, it is the
252b5132
RH
4177number of low-order zero bits the location counter must have after
4178advancement. For example @samp{.align 3} advances the location
4179counter until it a multiple of 8. If the location counter is already a
4180multiple of 8, no change is needed.
4181
4182This inconsistency is due to the different behaviors of the various
4183native assemblers for these systems which GAS must emulate.
4184GAS also provides @code{.balign} and @code{.p2align} directives,
4185described later, which have a consistent behavior across all
4186architectures (but are specific to GAS).
4187
ccf8a69b
BW
4188@node Altmacro
4189@section @code{.altmacro}
4190Enable alternate macro mode, enabling:
4191
4192@ftable @code
4193@item LOCAL @var{name} [ , @dots{} ]
4194One additional directive, @code{LOCAL}, is available. It is used to
4195generate a string replacement for each of the @var{name} arguments, and
4196replace any instances of @var{name} in each macro expansion. The
4197replacement string is unique in the assembly, and different for each
4198separate macro expansion. @code{LOCAL} allows you to write macros that
4199define symbols, without fear of conflict between separate macro expansions.
4200
4201@item String delimiters
4202You can write strings delimited in these other ways besides
4203@code{"@var{string}"}:
4204
4205@table @code
4206@item '@var{string}'
4207You can delimit strings with single-quote characters.
4208
4209@item <@var{string}>
4210You can delimit strings with matching angle brackets.
4211@end table
4212
4213@item single-character string escape
4214To include any single character literally in a string (even if the
4215character would otherwise have some special meaning), you can prefix the
4216character with @samp{!} (an exclamation mark). For example, you can
4217write @samp{<4.3 !> 5.4!!>} to get the literal text @samp{4.3 > 5.4!}.
4218
4219@item Expression results as strings
4220You can write @samp{%@var{expr}} to evaluate the expression @var{expr}
01642c12 4221and use the result as a string.
ccf8a69b
BW
4222@end ftable
4223
252b5132
RH
4224@node Ascii
4225@section @code{.ascii "@var{string}"}@dots{}
4226
4227@cindex @code{ascii} directive
4228@cindex string literals
4229@code{.ascii} expects zero or more string literals (@pxref{Strings})
4230separated by commas. It assembles each string (with no automatic
4231trailing zero byte) into consecutive addresses.
4232
4233@node Asciz
4234@section @code{.asciz "@var{string}"}@dots{}
4235
4236@cindex @code{asciz} directive
4237@cindex zero-terminated strings
4238@cindex null-terminated strings
4239@code{.asciz} is just like @code{.ascii}, but each string is followed by
4240a zero byte. The ``z'' in @samp{.asciz} stands for ``zero''.
4241
4242@node Balign
4243@section @code{.balign[wl] @var{abs-expr}, @var{abs-expr}, @var{abs-expr}}
4244
4245@cindex padding the location counter given number of bytes
4246@cindex @code{balign} directive
4247Pad the location counter (in the current subsection) to a particular
4248storage boundary. The first expression (which must be absolute) is the
4249alignment request in bytes. For example @samp{.balign 8} advances
4250the location counter until it is a multiple of 8. If the location counter
4251is already a multiple of 8, no change is needed.
4252
4253The second expression (also absolute) gives the fill value to be stored in the
4254padding bytes. It (and the comma) may be omitted. If it is omitted, the
4255padding bytes are normally zero. However, on some systems, if the section is
4256marked as containing code and the fill value is omitted, the space is filled
4257with no-op instructions.
4258
4259The third expression is also absolute, and is also optional. If it is present,
4260it is the maximum number of bytes that should be skipped by this alignment
4261directive. If doing the alignment would require skipping more bytes than the
4262specified maximum, then the alignment is not done at all. You can omit the
4263fill value (the second argument) entirely by simply using two commas after the
4264required alignment; this can be useful if you want the alignment to be filled
4265with no-op instructions when appropriate.
4266
4267@cindex @code{balignw} directive
4268@cindex @code{balignl} directive
4269The @code{.balignw} and @code{.balignl} directives are variants of the
4270@code{.balign} directive. The @code{.balignw} directive treats the fill
4271pattern as a two byte word value. The @code{.balignl} directives treats the
4272fill pattern as a four byte longword value. For example, @code{.balignw
42734,0x368d} will align to a multiple of 4. If it skips two bytes, they will be
4274filled in with the value 0x368d (the exact placement of the bytes depends upon
4275the endianness of the processor). If it skips 1 or 3 bytes, the fill value is
4276undefined.
4277
4278@node Byte
4279@section @code{.byte @var{expressions}}
4280
4281@cindex @code{byte} directive
4282@cindex integers, one byte
4283@code{.byte} expects zero or more expressions, separated by commas.
4284Each expression is assembled into the next byte.
4285
54cfded0 4286@node CFI directives
38462edf
JJ
4287@section @code{.cfi_sections @var{section_list}}
4288@cindex @code{cfi_sections} directive
4289@code{.cfi_sections} may be used to specify whether CFI directives
4290should emit @code{.eh_frame} section and/or @code{.debug_frame} section.
4291If @var{section_list} is @code{.eh_frame}, @code{.eh_frame} is emitted,
4292if @var{section_list} is @code{.debug_frame}, @code{.debug_frame} is emitted.
4293To emit both use @code{.eh_frame, .debug_frame}. The default if this
4294directive is not used is @code{.cfi_sections .eh_frame}.
4295
4b7d318b 4296@section @code{.cfi_startproc [simple]}
54cfded0
AM
4297@cindex @code{cfi_startproc} directive
4298@code{.cfi_startproc} is used at the beginning of each function that
4299should have an entry in @code{.eh_frame}. It initializes some internal
4b7d318b 4300data structures. Don't forget to close the function by
54cfded0
AM
4301@code{.cfi_endproc}.
4302
01642c12 4303Unless @code{.cfi_startproc} is used along with parameter @code{simple}
4b7d318b 4304it also emits some architecture dependent initial CFI instructions.
01642c12 4305
54cfded0
AM
4306@section @code{.cfi_endproc}
4307@cindex @code{cfi_endproc} directive
4308@code{.cfi_endproc} is used at the end of a function where it closes its
4309unwind entry previously opened by
b45619c0 4310@code{.cfi_startproc}, and emits it to @code{.eh_frame}.
54cfded0 4311
9b8ae42e
JJ
4312@section @code{.cfi_personality @var{encoding} [, @var{exp}]}
4313@code{.cfi_personality} defines personality routine and its encoding.
4314@var{encoding} must be a constant determining how the personality
4315should be encoded. If it is 255 (@code{DW_EH_PE_omit}), second
4316argument is not present, otherwise second argument should be
4317a constant or a symbol name. When using indirect encodings,
4318the symbol provided should be the location where personality
4319can be loaded from, not the personality routine itself.
4320The default after @code{.cfi_startproc} is @code{.cfi_personality 0xff},
4321no personality routine.
4322
4323@section @code{.cfi_lsda @var{encoding} [, @var{exp}]}
4324@code{.cfi_lsda} defines LSDA and its encoding.
4325@var{encoding} must be a constant determining how the LSDA
4326should be encoded. If it is 255 (@code{DW_EH_PE_omit}), second
4327argument is not present, otherwise second argument should be a constant
4328or a symbol name. The default after @code{.cfi_startproc} is @code{.cfi_lsda 0xff},
4329no LSDA.
4330
54cfded0 4331@section @code{.cfi_def_cfa @var{register}, @var{offset}}
01642c12 4332@code{.cfi_def_cfa} defines a rule for computing CFA as: @i{take
54cfded0
AM
4333address from @var{register} and add @var{offset} to it}.
4334
4335@section @code{.cfi_def_cfa_register @var{register}}
4336@code{.cfi_def_cfa_register} modifies a rule for computing CFA. From
4337now on @var{register} will be used instead of the old one. Offset
4338remains the same.
4339
4340@section @code{.cfi_def_cfa_offset @var{offset}}
4341@code{.cfi_def_cfa_offset} modifies a rule for computing CFA. Register
4342remains the same, but @var{offset} is new. Note that it is the
4343absolute offset that will be added to a defined register to compute
4344CFA address.
4345
4346@section @code{.cfi_adjust_cfa_offset @var{offset}}
4347Same as @code{.cfi_def_cfa_offset} but @var{offset} is a relative
4348value that is added/substracted from the previous offset.
4349
4350@section @code{.cfi_offset @var{register}, @var{offset}}
4351Previous value of @var{register} is saved at offset @var{offset} from
01642c12 4352CFA.
54cfded0 4353
17076204
RH
4354@section @code{.cfi_rel_offset @var{register}, @var{offset}}
4355Previous value of @var{register} is saved at offset @var{offset} from
4356the current CFA register. This is transformed to @code{.cfi_offset}
4357using the known displacement of the CFA register from the CFA.
4358This is often easier to use, because the number will match the
4359code it's annotating.
54cfded0 4360
4b7d318b
L
4361@section @code{.cfi_register @var{register1}, @var{register2}}
4362Previous value of @var{register1} is saved in register @var{register2}.
4363
4364@section @code{.cfi_restore @var{register}}
01642c12
RM
4365@code{.cfi_restore} says that the rule for @var{register} is now the
4366same as it was at the beginning of the function, after all initial
4b7d318b
L
4367instruction added by @code{.cfi_startproc} were executed.
4368
4369@section @code{.cfi_undefined @var{register}}
4370From now on the previous value of @var{register} can't be restored anymore.
4371
4372@section @code{.cfi_same_value @var{register}}
01642c12 4373Current value of @var{register} is the same like in the previous frame,
4b7d318b
L
4374i.e. no restoration needed.
4375
01642c12
RM
4376@section @code{.cfi_remember_state},
4377First save all current rules for all registers by @code{.cfi_remember_state},
4378then totally screw them up by subsequent @code{.cfi_*} directives and when
4379everything is hopelessly bad, use @code{.cfi_restore_state} to restore
4b7d318b
L
4380the previous saved state.
4381
4382@section @code{.cfi_return_column @var{register}}
01642c12 4383Change return column @var{register}, i.e. the return address is either
4b7d318b
L
4384directly in @var{register} or can be accessed by rules for @var{register}.
4385
63752a75
JJ
4386@section @code{.cfi_signal_frame}
4387Mark current function as signal trampoline.
4388
6749011b 4389@section @code{.cfi_window_save}
364b6d8b
JJ
4390SPARC register window has been saved.
4391
cdfbf930
RH
4392@section @code{.cfi_escape} @var{expression}[, @dots{}]
4393Allows the user to add arbitrary bytes to the unwind info. One
4394might use this to add OS-specific CFI opcodes, or generic CFI
4395opcodes that GAS does not yet support.
252b5132 4396
f1c4cc75
RH
4397@section @code{.cfi_val_encoded_addr @var{register}, @var{encoding}, @var{label}}
4398The current value of @var{register} is @var{label}. The value of @var{label}
4399will be encoded in the output file according to @var{encoding}; see the
4400description of @code{.cfi_personality} for details on this encoding.
4401
4402The usefulness of equating a register to a fixed label is probably
4403limited to the return address register. Here, it can be useful to
4404mark a code segment that has only one return address which is reached
4405by a direct branch and no copy of the return address exists in memory
4406or another register.
4407
ccf8a69b
BW
4408@node Comm
4409@section @code{.comm @var{symbol} , @var{length} }
bd0eb99b 4410
ccf8a69b
BW
4411@cindex @code{comm} directive
4412@cindex symbol, common
4413@code{.comm} declares a common symbol named @var{symbol}. When linking, a
4414common symbol in one object file may be merged with a defined or common symbol
4415of the same name in another object file. If @code{@value{LD}} does not see a
4416definition for the symbol--just one or more common symbols--then it will
4417allocate @var{length} bytes of uninitialized memory. @var{length} must be an
4418absolute expression. If @code{@value{LD}} sees multiple common symbols with
4419the same name, and they do not all have the same size, it will allocate space
4420using the largest size.
07a53e5c 4421
c1711530
DK
4422@ifset COFF-ELF
4423When using ELF or (as a GNU extension) PE, the @code{.comm} directive takes
01642c12 4424an optional third argument. This is the desired alignment of the symbol,
c1711530
DK
4425specified for ELF as a byte boundary (for example, an alignment of 16 means
4426that the least significant 4 bits of the address should be zero), and for PE
4427as a power of two (for example, an alignment of 5 means aligned to a 32-byte
01642c12 4428boundary). The alignment must be an absolute expression, and it must be a
c1711530 4429power of two. If @code{@value{LD}} allocates uninitialized memory for the
01642c12 4430common symbol, it will use the alignment when placing the symbol. If no
c1711530 4431alignment is specified, @command{@value{AS}} will set the alignment to the
ccf8a69b 4432largest power of two less than or equal to the size of the symbol, up to a
c1711530
DK
4433maximum of 16 on ELF, or the default section alignment of 4 on PE@footnote{This
4434is not the same as the executable image file alignment controlled by @code{@value{LD}}'s
4435@samp{--section-alignment} option; image file sections in PE are aligned to
4436multiples of 4096, which is far too large an alignment for ordinary variables.
4437It is rather the default alignment for (non-debug) sections within object
4438(@samp{*.o}) files, which are less strictly aligned.}.
ccf8a69b 4439@end ifset
cd1fcb49 4440
ccf8a69b
BW
4441@ifset HPPA
4442The syntax for @code{.comm} differs slightly on the HPPA. The syntax is
4443@samp{@var{symbol} .comm, @var{length}}; @var{symbol} is optional.
4444@end ifset
07a53e5c 4445
252b5132
RH
4446@node Data
4447@section @code{.data @var{subsection}}
4448
4449@cindex @code{data} directive
a4fb0134 4450@code{.data} tells @command{@value{AS}} to assemble the following statements onto the
252b5132
RH
4451end of the data subsection numbered @var{subsection} (which is an
4452absolute expression). If @var{subsection} is omitted, it defaults
4453to zero.
4454
4455@ifset COFF
4456@node Def
4457@section @code{.def @var{name}}
4458
4459@cindex @code{def} directive
4460@cindex COFF symbols, debugging
4461@cindex debugging COFF symbols
4462Begin defining debugging information for a symbol @var{name}; the
4463definition extends until the @code{.endef} directive is encountered.
4464@ifset BOUT
4465
a4fb0134 4466This directive is only observed when @command{@value{AS}} is configured for COFF
252b5132
RH
4467format output; when producing @code{b.out}, @samp{.def} is recognized,
4468but ignored.
4469@end ifset
4470@end ifset
4471
4472@ifset aout-bout
4473@node Desc
4474@section @code{.desc @var{symbol}, @var{abs-expression}}
4475
4476@cindex @code{desc} directive
4477@cindex COFF symbol descriptor
4478@cindex symbol descriptor, COFF
4479This directive sets the descriptor of the symbol (@pxref{Symbol Attributes})
4480to the low 16 bits of an absolute expression.
4481
4482@ifset COFF
a4fb0134 4483The @samp{.desc} directive is not available when @command{@value{AS}} is
252b5132 4484configured for COFF output; it is only for @code{a.out} or @code{b.out}
a4fb0134 4485object format. For the sake of compatibility, @command{@value{AS}} accepts
252b5132
RH
4486it, but produces no output, when configured for COFF.
4487@end ifset
4488@end ifset
4489
4490@ifset COFF
4491@node Dim
4492@section @code{.dim}
4493
4494@cindex @code{dim} directive
4495@cindex COFF auxiliary symbol information
4496@cindex auxiliary symbol information, COFF
4497This directive is generated by compilers to include auxiliary debugging
4498information in the symbol table. It is only permitted inside
4499@code{.def}/@code{.endef} pairs.
4500@ifset BOUT
4501
4502@samp{.dim} is only meaningful when generating COFF format output; when
a4fb0134 4503@command{@value{AS}} is generating @code{b.out}, it accepts this directive but
252b5132
RH
4504ignores it.
4505@end ifset
4506@end ifset
4507
4508@node Double
4509@section @code{.double @var{flonums}}
4510
4511@cindex @code{double} directive
4512@cindex floating point numbers (double)
4513@code{.double} expects zero or more flonums, separated by commas. It
4514assembles floating point numbers.
4515@ifset GENERIC
4516The exact kind of floating point numbers emitted depends on how
a4fb0134 4517@command{@value{AS}} is configured. @xref{Machine Dependencies}.
252b5132
RH
4518@end ifset
4519@ifclear GENERIC
4520@ifset IEEEFLOAT
4521On the @value{TARGET} family @samp{.double} emits 64-bit floating-point numbers
4522in @sc{ieee} format.
4523@end ifset
4524@end ifclear
4525
4526@node Eject
4527@section @code{.eject}
4528
4529@cindex @code{eject} directive
4530@cindex new page, in listings
4531@cindex page, in listings
4532@cindex listing control: new page
4533Force a page break at this point, when generating assembly listings.
4534
4535@node Else
4536@section @code{.else}
4537
4538@cindex @code{else} directive
a4fb0134 4539@code{.else} is part of the @command{@value{AS}} support for conditional
96e9638b 4540assembly; see @ref{If,,@code{.if}}. It marks the beginning of a section
252b5132
RH
4541of code to be assembled if the condition for the preceding @code{.if}
4542was false.
4543
3fd9f047
TW
4544@node Elseif
4545@section @code{.elseif}
4546
4547@cindex @code{elseif} directive
a4fb0134 4548@code{.elseif} is part of the @command{@value{AS}} support for conditional
96e9638b 4549assembly; see @ref{If,,@code{.if}}. It is shorthand for beginning a new
3fd9f047
TW
4550@code{.if} block that would otherwise fill the entire @code{.else} section.
4551
252b5132
RH
4552@node End
4553@section @code{.end}
4554
4555@cindex @code{end} directive
a4fb0134 4556@code{.end} marks the end of the assembly file. @command{@value{AS}} does not
252b5132
RH
4557process anything in the file past the @code{.end} directive.
4558
4559@ifset COFF
4560@node Endef
4561@section @code{.endef}
4562
4563@cindex @code{endef} directive
4564This directive flags the end of a symbol definition begun with
4565@code{.def}.
4566@ifset BOUT
4567
4568@samp{.endef} is only meaningful when generating COFF format output; if
a4fb0134 4569@command{@value{AS}} is configured to generate @code{b.out}, it accepts this
252b5132
RH
4570directive but ignores it.
4571@end ifset
4572@end ifset
4573
4574@node Endfunc
4575@section @code{.endfunc}
4576@cindex @code{endfunc} directive
4577@code{.endfunc} marks the end of a function specified with @code{.func}.
4578
4579@node Endif
4580@section @code{.endif}
4581
4582@cindex @code{endif} directive
a4fb0134 4583@code{.endif} is part of the @command{@value{AS}} support for conditional assembly;
252b5132
RH
4584it marks the end of a block of code that is only assembled
4585conditionally. @xref{If,,@code{.if}}.
4586
4587@node Equ
4588@section @code{.equ @var{symbol}, @var{expression}}
4589
4590@cindex @code{equ} directive
4591@cindex assigning values to symbols
4592@cindex symbols, assigning values to
4593This directive sets the value of @var{symbol} to @var{expression}.
96e9638b 4594It is synonymous with @samp{.set}; see @ref{Set,,@code{.set}}.
252b5132
RH
4595
4596@ifset HPPA
01642c12 4597The syntax for @code{equ} on the HPPA is
252b5132
RH
4598@samp{@var{symbol} .equ @var{expression}}.
4599@end ifset
4600
3c9b82ba 4601@ifset Z80
01642c12
RM
4602The syntax for @code{equ} on the Z80 is
4603@samp{@var{symbol} equ @var{expression}}.
3c9b82ba 4604On the Z80 it is an eror if @var{symbol} is already defined,
01642c12 4605but the symbol is not protected from later redefinition.
96e9638b 4606Compare @ref{Equiv}.
3c9b82ba
NC
4607@end ifset
4608
252b5132
RH
4609@node Equiv
4610@section @code{.equiv @var{symbol}, @var{expression}}
4611@cindex @code{equiv} directive
4612The @code{.equiv} directive is like @code{.equ} and @code{.set}, except that
8dfa0188
NC
4613the assembler will signal an error if @var{symbol} is already defined. Note a
4614symbol which has been referenced but not actually defined is considered to be
4615undefined.
252b5132 4616
01642c12 4617Except for the contents of the error message, this is roughly equivalent to
252b5132
RH
4618@smallexample
4619.ifdef SYM
4620.err
4621.endif
4622.equ SYM,VAL
4623@end smallexample
9497f5ac
NC
4624plus it protects the symbol from later redefinition.
4625
4626@node Eqv
4627@section @code{.eqv @var{symbol}, @var{expression}}
4628@cindex @code{eqv} directive
4629The @code{.eqv} directive is like @code{.equiv}, but no attempt is made to
4630evaluate the expression or any part of it immediately. Instead each time
4631the resulting symbol is used in an expression, a snapshot of its current
4632value is taken.
252b5132
RH
4633
4634@node Err
4635@section @code{.err}
4636@cindex @code{err} directive
a4fb0134
SC
4637If @command{@value{AS}} assembles a @code{.err} directive, it will print an error
4638message and, unless the @option{-Z} option was used, it will not generate an
f9eb6721 4639object file. This can be used to signal an error in conditionally compiled code.
252b5132 4640
d190d046
HPN
4641@node Error
4642@section @code{.error "@var{string}"}
4643@cindex error directive
4644
4645Similarly to @code{.err}, this directive emits an error, but you can specify a
4646string that will be emitted as the error message. If you don't specify the
4647message, it defaults to @code{".error directive invoked in source file"}.
4648@xref{Errors, ,Error and Warning Messages}.
4649
4650@smallexample
4651 .error "This code has not been assembled and tested."
4652@end smallexample
4653
252b5132
RH
4654@node Exitm
4655@section @code{.exitm}
4656Exit early from the current macro definition. @xref{Macro}.
4657
4658@node Extern
4659@section @code{.extern}
4660
4661@cindex @code{extern} directive
4662@code{.extern} is accepted in the source program---for compatibility
a4fb0134 4663with other assemblers---but it is ignored. @command{@value{AS}} treats
252b5132
RH
4664all undefined symbols as external.
4665
4666@node Fail
4667@section @code{.fail @var{expression}}
4668
4669@cindex @code{fail} directive
4670Generates an error or a warning. If the value of the @var{expression} is 500
a4fb0134
SC
4671or more, @command{@value{AS}} will print a warning message. If the value is less
4672than 500, @command{@value{AS}} will print an error message. The message will
252b5132
RH
4673include the value of @var{expression}. This can occasionally be useful inside
4674complex nested macros or conditional assembly.
4675
252b5132 4676@node File
14082c76 4677@section @code{.file}
252b5132 4678@cindex @code{file} directive
14082c76
BW
4679
4680@ifclear no-file-dir
4681There are two different versions of the @code{.file} directive. Targets
4682that support DWARF2 line number information use the DWARF2 version of
4683@code{.file}. Other targets use the default version.
4684
4685@subheading Default Version
4686
252b5132
RH
4687@cindex logical file name
4688@cindex file name, logical
14082c76
BW
4689This version of the @code{.file} directive tells @command{@value{AS}} that we
4690are about to start a new logical file. The syntax is:
4691
4692@smallexample
4693.file @var{string}
4694@end smallexample
4695
4696@var{string} is the new file name. In general, the filename is
252b5132
RH
4697recognized whether or not it is surrounded by quotes @samp{"}; but if you wish
4698to specify an empty file name, you must give the quotes--@code{""}. This
4699statement may go away in future: it is only recognized to be compatible with
a4fb0134 4700old @command{@value{AS}} programs.
14082c76
BW
4701
4702@subheading DWARF2 Version
252b5132
RH
4703@end ifclear
4704
14082c76
BW
4705When emitting DWARF2 line number information, @code{.file} assigns filenames
4706to the @code{.debug_line} file name table. The syntax is:
4707
4708@smallexample
4709.file @var{fileno} @var{filename}
4710@end smallexample
4711
4712The @var{fileno} operand should be a unique positive integer to use as the
4713index of the entry in the table. The @var{filename} operand is a C string
4714literal.
4715
4716The detail of filename indices is exposed to the user because the filename
4717table is shared with the @code{.debug_info} section of the DWARF2 debugging
4718information, and thus the user must know the exact indices that table
4719entries will have.
4720
252b5132
RH
4721@node Fill
4722@section @code{.fill @var{repeat} , @var{size} , @var{value}}
4723
4724@cindex @code{fill} directive
4725@cindex writing patterns in memory
4726@cindex patterns, writing in memory
bc64be0c 4727@var{repeat}, @var{size} and @var{value} are absolute expressions.
252b5132
RH
4728This emits @var{repeat} copies of @var{size} bytes. @var{Repeat}
4729may be zero or more. @var{Size} may be zero or more, but if it is
4730more than 8, then it is deemed to have the value 8, compatible with
4731other people's assemblers. The contents of each @var{repeat} bytes
4732is taken from an 8-byte number. The highest order 4 bytes are
4733zero. The lowest order 4 bytes are @var{value} rendered in the
a4fb0134 4734byte-order of an integer on the computer @command{@value{AS}} is assembling for.
252b5132
RH
4735Each @var{size} bytes in a repetition is taken from the lowest order
4736@var{size} bytes of this number. Again, this bizarre behavior is
4737compatible with other people's assemblers.
4738
4739@var{size} and @var{value} are optional.
4740If the second comma and @var{value} are absent, @var{value} is
4741assumed zero. If the first comma and following tokens are absent,
4742@var{size} is assumed to be 1.
4743
4744@node Float
4745@section @code{.float @var{flonums}}
4746
4747@cindex floating point numbers (single)
4748@cindex @code{float} directive
4749This directive assembles zero or more flonums, separated by commas. It
4750has the same effect as @code{.single}.
4751@ifset GENERIC
4752The exact kind of floating point numbers emitted depends on how
a4fb0134 4753@command{@value{AS}} is configured.
252b5132
RH
4754@xref{Machine Dependencies}.
4755@end ifset
4756@ifclear GENERIC
4757@ifset IEEEFLOAT
4758On the @value{TARGET} family, @code{.float} emits 32-bit floating point numbers
4759in @sc{ieee} format.
4760@end ifset
4761@end ifclear
4762
4763@node Func
4764@section @code{.func @var{name}[,@var{label}]}
4765@cindex @code{func} directive
4766@code{.func} emits debugging information to denote function @var{name}, and
4767is ignored unless the file is assembled with debugging enabled.
05da4302 4768Only @samp{--gstabs[+]} is currently supported.
252b5132
RH
4769@var{label} is the entry point of the function and if omitted @var{name}
4770prepended with the @samp{leading char} is used.
4771@samp{leading char} is usually @code{_} or nothing, depending on the target.
4772All functions are currently defined to have @code{void} return type.
4773The function must be terminated with @code{.endfunc}.
4774
4775@node Global
4776@section @code{.global @var{symbol}}, @code{.globl @var{symbol}}
4777
4778@cindex @code{global} directive
4779@cindex symbol, making visible to linker
4780@code{.global} makes the symbol visible to @code{@value{LD}}. If you define
4781@var{symbol} in your partial program, its value is made available to
4782other partial programs that are linked with it. Otherwise,
4783@var{symbol} takes its attributes from a symbol of the same name
4784from another file linked into the same program.
4785
4786Both spellings (@samp{.globl} and @samp{.global}) are accepted, for
4787compatibility with other assemblers.
4788
4789@ifset HPPA
4790On the HPPA, @code{.global} is not always enough to make it accessible to other
4791partial programs. You may need the HPPA-only @code{.EXPORT} directive as well.
96e9638b 4792@xref{HPPA Directives, ,HPPA Assembler Directives}.
252b5132
RH
4793@end ifset
4794
c91d2e08 4795@ifset ELF
3a99f02f
DJ
4796@node Gnu_attribute
4797@section @code{.gnu_attribute @var{tag},@var{value}}
4798Record a @sc{gnu} object attribute for this file. @xref{Object Attributes}.
4799
c91d2e08
NC
4800@node Hidden
4801@section @code{.hidden @var{names}}
4802
c1253627
NC
4803@cindex @code{hidden} directive
4804@cindex visibility
ed9589d4 4805This is one of the ELF visibility directives. The other two are
01642c12 4806@code{.internal} (@pxref{Internal,,@code{.internal}}) and
a349d9dd 4807@code{.protected} (@pxref{Protected,,@code{.protected}}).
c91d2e08
NC
4808
4809This directive overrides the named symbols default visibility (which is set by
4810their binding: local, global or weak). The directive sets the visibility to
4811@code{hidden} which means that the symbols are not visible to other components.
01642c12 4812Such symbols are always considered to be @code{protected} as well.
c91d2e08
NC
4813@end ifset
4814
252b5132
RH
4815@node hword
4816@section @code{.hword @var{expressions}}
4817
4818@cindex @code{hword} directive
4819@cindex integers, 16-bit
4820@cindex numbers, 16-bit
4821@cindex sixteen bit integers
4822This expects zero or more @var{expressions}, and emits
4823a 16 bit number for each.
4824
4825@ifset GENERIC
4826This directive is a synonym for @samp{.short}; depending on the target
4827architecture, it may also be a synonym for @samp{.word}.
4828@end ifset
4829@ifclear GENERIC
4830@ifset W32
4831This directive is a synonym for @samp{.short}.
4832@end ifset
4833@ifset W16
4834This directive is a synonym for both @samp{.short} and @samp{.word}.
4835@end ifset
4836@end ifclear
4837
4838@node Ident
4839@section @code{.ident}
4840
4841@cindex @code{ident} directive
cb4c78d6
BE
4842
4843This directive is used by some assemblers to place tags in object files. The
4844behavior of this directive varies depending on the target. When using the
4845a.out object file format, @command{@value{AS}} simply accepts the directive for
4846source-file compatibility with existing assemblers, but does not emit anything
4847for it. When using COFF, comments are emitted to the @code{.comment} or
4848@code{.rdata} section, depending on the target. When using ELF, comments are
4849emitted to the @code{.comment} section.
252b5132
RH
4850
4851@node If
4852@section @code{.if @var{absolute expression}}
4853
4854@cindex conditional assembly
4855@cindex @code{if} directive
4856@code{.if} marks the beginning of a section of code which is only
4857considered part of the source program being assembled if the argument
4858(which must be an @var{absolute expression}) is non-zero. The end of
4859the conditional section of code must be marked by @code{.endif}
4860(@pxref{Endif,,@code{.endif}}); optionally, you may include code for the
4861alternative condition, flagged by @code{.else} (@pxref{Else,,@code{.else}}).
3fd9f047
TW
4862If you have several conditions to check, @code{.elseif} may be used to avoid
4863nesting blocks if/else within each subsequent @code{.else} block.
252b5132
RH
4864
4865The following variants of @code{.if} are also supported:
4866@table @code
4867@cindex @code{ifdef} directive
4868@item .ifdef @var{symbol}
4869Assembles the following section of code if the specified @var{symbol}
8dfa0188
NC
4870has been defined. Note a symbol which has been referenced but not yet defined
4871is considered to be undefined.
252b5132 4872
26aca5f6
JB
4873@cindex @code{ifb} directive
4874@item .ifb @var{text}
4875Assembles the following section of code if the operand is blank (empty).
4876
252b5132
RH
4877@cindex @code{ifc} directive
4878@item .ifc @var{string1},@var{string2}
4879Assembles the following section of code if the two strings are the same. The
4880strings may be optionally quoted with single quotes. If they are not quoted,
4881the first string stops at the first comma, and the second string stops at the
4882end of the line. Strings which contain whitespace should be quoted. The
4883string comparison is case sensitive.
4884
4885@cindex @code{ifeq} directive
4886@item .ifeq @var{absolute expression}
4887Assembles the following section of code if the argument is zero.
4888
4889@cindex @code{ifeqs} directive
4890@item .ifeqs @var{string1},@var{string2}
4891Another form of @code{.ifc}. The strings must be quoted using double quotes.
4892
4893@cindex @code{ifge} directive
4894@item .ifge @var{absolute expression}
4895Assembles the following section of code if the argument is greater than or
4896equal to zero.
4897
4898@cindex @code{ifgt} directive
4899@item .ifgt @var{absolute expression}
4900Assembles the following section of code if the argument is greater than zero.
4901
4902@cindex @code{ifle} directive
4903@item .ifle @var{absolute expression}
4904Assembles the following section of code if the argument is less than or equal
4905to zero.
4906
4907@cindex @code{iflt} directive
4908@item .iflt @var{absolute expression}
4909Assembles the following section of code if the argument is less than zero.
4910
26aca5f6
JB
4911@cindex @code{ifnb} directive
4912@item .ifnb @var{text}
4913Like @code{.ifb}, but the sense of the test is reversed: this assembles the
4914following section of code if the operand is non-blank (non-empty).
4915
252b5132
RH
4916@cindex @code{ifnc} directive
4917@item .ifnc @var{string1},@var{string2}.
4918Like @code{.ifc}, but the sense of the test is reversed: this assembles the
4919following section of code if the two strings are not the same.
4920
4921@cindex @code{ifndef} directive
4922@cindex @code{ifnotdef} directive
4923@item .ifndef @var{symbol}
4924@itemx .ifnotdef @var{symbol}
4925Assembles the following section of code if the specified @var{symbol}
8dfa0188
NC
4926has not been defined. Both spelling variants are equivalent. Note a symbol
4927which has been referenced but not yet defined is considered to be undefined.
252b5132
RH
4928
4929@cindex @code{ifne} directive
4930@item .ifne @var{absolute expression}
4931Assembles the following section of code if the argument is not equal to zero
4932(in other words, this is equivalent to @code{.if}).
4933
4934@cindex @code{ifnes} directive
4935@item .ifnes @var{string1},@var{string2}
4936Like @code{.ifeqs}, but the sense of the test is reversed: this assembles the
4937following section of code if the two strings are not the same.
4938@end table
4939
7e005732
NC
4940@node Incbin
4941@section @code{.incbin "@var{file}"[,@var{skip}[,@var{count}]]}
4942
4943@cindex @code{incbin} directive
4944@cindex binary files, including
4945The @code{incbin} directive includes @var{file} verbatim at the current
4946location. You can control the search paths used with the @samp{-I} command-line
4947option (@pxref{Invoking,,Command-Line Options}). Quotation marks are required
4948around @var{file}.
4949
4950The @var{skip} argument skips a number of bytes from the start of the
4951@var{file}. The @var{count} argument indicates the maximum number of bytes to
15dcfbc3
NC
4952read. Note that the data is not aligned in any way, so it is the user's
4953responsibility to make sure that proper alignment is provided both before and
4954after the @code{incbin} directive.
7e005732 4955
252b5132
RH
4956@node Include
4957@section @code{.include "@var{file}"}
4958
4959@cindex @code{include} directive
4960@cindex supporting files, including
4961@cindex files, including
4962This directive provides a way to include supporting files at specified
4963points in your source program. The code from @var{file} is assembled as
4964if it followed the point of the @code{.include}; when the end of the
4965included file is reached, assembly of the original file continues. You
4966can control the search paths used with the @samp{-I} command-line option
4967(@pxref{Invoking,,Command-Line Options}). Quotation marks are required
4968around @var{file}.
4969
4970@node Int
4971@section @code{.int @var{expressions}}
4972
4973@cindex @code{int} directive
4974@cindex integers, 32-bit
4975Expect zero or more @var{expressions}, of any section, separated by commas.
4976For each expression, emit a number that, at run time, is the value of that
4977expression. The byte order and bit size of the number depends on what kind
4978of target the assembly is for.
4979
4980@ifclear GENERIC
4981@ifset H8
7be1c489 4982On most forms of the H8/300, @code{.int} emits 16-bit
c2dcd04e 4983integers. On the H8/300H and the Renesas SH, however, @code{.int} emits
252b5132
RH
498432-bit integers.
4985@end ifset
4986@end ifclear
4987
c91d2e08
NC
4988@ifset ELF
4989@node Internal
4990@section @code{.internal @var{names}}
4991
c1253627
NC
4992@cindex @code{internal} directive
4993@cindex visibility
ed9589d4 4994This is one of the ELF visibility directives. The other two are
01642c12 4995@code{.hidden} (@pxref{Hidden,,@code{.hidden}}) and
a349d9dd 4996@code{.protected} (@pxref{Protected,,@code{.protected}}).
c91d2e08
NC
4997
4998This directive overrides the named symbols default visibility (which is set by
4999their binding: local, global or weak). The directive sets the visibility to
5000@code{internal} which means that the symbols are considered to be @code{hidden}
c1253627 5001(i.e., not visible to other components), and that some extra, processor specific
c91d2e08
NC
5002processing must also be performed upon the symbols as well.
5003@end ifset
5004
252b5132
RH
5005@node Irp
5006@section @code{.irp @var{symbol},@var{values}}@dots{}
5007
5008@cindex @code{irp} directive
5009Evaluate a sequence of statements assigning different values to @var{symbol}.
5010The sequence of statements starts at the @code{.irp} directive, and is
5011terminated by an @code{.endr} directive. For each @var{value}, @var{symbol} is
5012set to @var{value}, and the sequence of statements is assembled. If no
5013@var{value} is listed, the sequence of statements is assembled once, with
5014@var{symbol} set to the null string. To refer to @var{symbol} within the
5015sequence of statements, use @var{\symbol}.
5016
5017For example, assembling
5018
5019@example
5020 .irp param,1,2,3
5021 move d\param,sp@@-
5022 .endr
5023@end example
5024
5025is equivalent to assembling
5026
5027@example
5028 move d1,sp@@-
5029 move d2,sp@@-
5030 move d3,sp@@-
5031@end example
5032
96e9638b 5033For some caveats with the spelling of @var{symbol}, see also @ref{Macro}.
5e75c3ab 5034
252b5132
RH
5035@node Irpc
5036@section @code{.irpc @var{symbol},@var{values}}@dots{}
5037
5038@cindex @code{irpc} directive
5039Evaluate a sequence of statements assigning different values to @var{symbol}.
5040The sequence of statements starts at the @code{.irpc} directive, and is
5041terminated by an @code{.endr} directive. For each character in @var{value},
5042@var{symbol} is set to the character, and the sequence of statements is
5043assembled. If no @var{value} is listed, the sequence of statements is
5044assembled once, with @var{symbol} set to the null string. To refer to
5045@var{symbol} within the sequence of statements, use @var{\symbol}.
5046
5047For example, assembling
5048
5049@example
5050 .irpc param,123
5051 move d\param,sp@@-
5052 .endr
5053@end example
5054
5055is equivalent to assembling
5056
5057@example
5058 move d1,sp@@-
5059 move d2,sp@@-
5060 move d3,sp@@-
5061@end example
5062
5e75c3ab
JB
5063For some caveats with the spelling of @var{symbol}, see also the discussion
5064at @xref{Macro}.
5065
252b5132
RH
5066@node Lcomm
5067@section @code{.lcomm @var{symbol} , @var{length}}
5068
5069@cindex @code{lcomm} directive
5070@cindex local common symbols
5071@cindex symbols, local common
5072Reserve @var{length} (an absolute expression) bytes for a local common
5073denoted by @var{symbol}. The section and value of @var{symbol} are
5074those of the new local common. The addresses are allocated in the bss
5075section, so that at run-time the bytes start off zeroed. @var{Symbol}
5076is not declared global (@pxref{Global,,@code{.global}}), so is normally
5077not visible to @code{@value{LD}}.
5078
5079@ifset GENERIC
5080Some targets permit a third argument to be used with @code{.lcomm}. This
5081argument specifies the desired alignment of the symbol in the bss section.
5082@end ifset
5083
5084@ifset HPPA
5085The syntax for @code{.lcomm} differs slightly on the HPPA. The syntax is
5086@samp{@var{symbol} .lcomm, @var{length}}; @var{symbol} is optional.
5087@end ifset
5088
5089@node Lflags
5090@section @code{.lflags}
5091
5092@cindex @code{lflags} directive (ignored)
a4fb0134 5093@command{@value{AS}} accepts this directive, for compatibility with other
252b5132
RH
5094assemblers, but ignores it.
5095
5096@ifclear no-line-dir
5097@node Line
5098@section @code{.line @var{line-number}}
5099
5100@cindex @code{line} directive
252b5132
RH
5101@cindex logical line number
5102@ifset aout-bout
5103Change the logical line number. @var{line-number} must be an absolute
5104expression. The next line has that logical line number. Therefore any other
5105statements on the current line (after a statement separator character) are
5106reported as on logical line number @var{line-number} @minus{} 1. One day
a4fb0134 5107@command{@value{AS}} will no longer support this directive: it is recognized only
252b5132 5108for compatibility with existing assembler programs.
252b5132
RH
5109@end ifset
5110
252b5132 5111Even though this is a directive associated with the @code{a.out} or
a4fb0134 5112@code{b.out} object-code formats, @command{@value{AS}} still recognizes it
252b5132
RH
5113when producing COFF output, and treats @samp{.line} as though it
5114were the COFF @samp{.ln} @emph{if} it is found outside a
5115@code{.def}/@code{.endef} pair.
5116
5117Inside a @code{.def}, @samp{.line} is, instead, one of the directives
5118used by compilers to generate auxiliary symbol information for
5119debugging.
5120@end ifclear
5121
5122@node Linkonce
5123@section @code{.linkonce [@var{type}]}
5124@cindex COMDAT
5125@cindex @code{linkonce} directive
5126@cindex common sections
5127Mark the current section so that the linker only includes a single copy of it.
5128This may be used to include the same section in several different object files,
5129but ensure that the linker will only include it once in the final output file.
5130The @code{.linkonce} pseudo-op must be used for each instance of the section.
5131Duplicate sections are detected based on the section name, so it should be
5132unique.
5133
5134This directive is only supported by a few object file formats; as of this
5135writing, the only object file format which supports it is the Portable
5136Executable format used on Windows NT.
5137
5138The @var{type} argument is optional. If specified, it must be one of the
5139following strings. For example:
5140@smallexample
5141.linkonce same_size
5142@end smallexample
5143Not all types may be supported on all object file formats.
5144
5145@table @code
5146@item discard
5147Silently discard duplicate sections. This is the default.
5148
5149@item one_only
5150Warn if there are duplicate sections, but still keep only one copy.
5151
5152@item same_size
5153Warn if any of the duplicates have different sizes.
5154
5155@item same_contents
5156Warn if any of the duplicates do not have exactly the same contents.
5157@end table
5158
ccf8a69b
BW
5159@node List
5160@section @code{.list}
5161
5162@cindex @code{list} directive
5163@cindex listing control, turning on
5164Control (in conjunction with the @code{.nolist} directive) whether or
5165not assembly listings are generated. These two directives maintain an
5166internal counter (which is zero initially). @code{.list} increments the
5167counter, and @code{.nolist} decrements it. Assembly listings are
5168generated whenever the counter is greater than zero.
5169
5170By default, listings are disabled. When you enable them (with the
5171@samp{-a} command line option; @pxref{Invoking,,Command-Line Options}),
5172the initial value of the listing counter is one.
5173
252b5132
RH
5174@node Ln
5175@section @code{.ln @var{line-number}}
5176
5177@cindex @code{ln} directive
5178@ifclear no-line-dir
5179@samp{.ln} is a synonym for @samp{.line}.
5180@end ifclear
5181@ifset no-line-dir
a4fb0134 5182Tell @command{@value{AS}} to change the logical line number. @var{line-number}
252b5132
RH
5183must be an absolute expression. The next line has that logical
5184line number, so any other statements on the current line (after a
5185statement separator character @code{;}) are reported as on logical
5186line number @var{line-number} @minus{} 1.
5187@ifset BOUT
5188
a4fb0134 5189This directive is accepted, but ignored, when @command{@value{AS}} is
252b5132
RH
5190configured for @code{b.out}; its effect is only associated with COFF
5191output format.
5192@end ifset
5193@end ifset
5194
ccf8a69b
BW
5195@node Loc
5196@section @code{.loc @var{fileno} @var{lineno} [@var{column}] [@var{options}]}
5197@cindex @code{loc} directive
5198When emitting DWARF2 line number information,
5199the @code{.loc} directive will add a row to the @code{.debug_line} line
5200number matrix corresponding to the immediately following assembly
5201instruction. The @var{fileno}, @var{lineno}, and optional @var{column}
5202arguments will be applied to the @code{.debug_line} state machine before
5203the row is added.
252b5132 5204
ccf8a69b
BW
5205The @var{options} are a sequence of the following tokens in any order:
5206
5207@table @code
5208@item basic_block
5209This option will set the @code{basic_block} register in the
5210@code{.debug_line} state machine to @code{true}.
5211
5212@item prologue_end
5213This option will set the @code{prologue_end} register in the
5214@code{.debug_line} state machine to @code{true}.
5215
5216@item epilogue_begin
5217This option will set the @code{epilogue_begin} register in the
5218@code{.debug_line} state machine to @code{true}.
5219
5220@item is_stmt @var{value}
5221This option will set the @code{is_stmt} register in the
01642c12 5222@code{.debug_line} state machine to @code{value}, which must be
ccf8a69b
BW
5223either 0 or 1.
5224
5225@item isa @var{value}
5226This directive will set the @code{isa} register in the @code{.debug_line}
5227state machine to @var{value}, which must be an unsigned integer.
5228
92846e72
CC
5229@item discriminator @var{value}
5230This directive will set the @code{discriminator} register in the @code{.debug_line}
5231state machine to @var{value}, which must be an unsigned integer.
5232
ccf8a69b
BW
5233@end table
5234
5235@node Loc_mark_labels
5236@section @code{.loc_mark_labels @var{enable}}
5237@cindex @code{loc_mark_labels} directive
5238When emitting DWARF2 line number information,
5239the @code{.loc_mark_labels} directive makes the assembler emit an entry
5240to the @code{.debug_line} line number matrix with the @code{basic_block}
5241register in the state machine set whenever a code label is seen.
5242The @var{enable} argument should be either 1 or 0, to enable or disable
5243this function respectively.
252b5132 5244
4d4175af
BW
5245@ifset ELF
5246@node Local
5247@section @code{.local @var{names}}
5248
5249@cindex @code{local} directive
5250This directive, which is available for ELF targets, marks each symbol in
5251the comma-separated list of @code{names} as a local symbol so that it
5252will not be externally visible. If the symbols do not already exist,
5253they will be created.
5254
5255For targets where the @code{.lcomm} directive (@pxref{Lcomm}) does not
5256accept an alignment argument, which is the case for most ELF targets,
5257the @code{.local} directive can be used in combination with @code{.comm}
5258(@pxref{Comm}) to define aligned local common data.
5259@end ifset
5260
252b5132
RH
5261@node Long
5262@section @code{.long @var{expressions}}
5263
5264@cindex @code{long} directive
96e9638b 5265@code{.long} is the same as @samp{.int}. @xref{Int,,@code{.int}}.
252b5132
RH
5266
5267@ignore
5268@c no one seems to know what this is for or whether this description is
5269@c what it really ought to do
5270@node Lsym
5271@section @code{.lsym @var{symbol}, @var{expression}}
5272
5273@cindex @code{lsym} directive
5274@cindex symbol, not referenced in assembly
5275@code{.lsym} creates a new symbol named @var{symbol}, but does not put it in
5276the hash table, ensuring it cannot be referenced by name during the
5277rest of the assembly. This sets the attributes of the symbol to be
5278the same as the expression value:
5279@smallexample
5280@var{other} = @var{descriptor} = 0
5281@var{type} = @r{(section of @var{expression})}
5282@var{value} = @var{expression}
5283@end smallexample
5284@noindent
5285The new symbol is not flagged as external.
5286@end ignore
5287
5288@node Macro
5289@section @code{.macro}
5290
5291@cindex macros
5292The commands @code{.macro} and @code{.endm} allow you to define macros that
5293generate assembly output. For example, this definition specifies a macro
5294@code{sum} that puts a sequence of numbers into memory:
5295
5296@example
5297 .macro sum from=0, to=5
5298 .long \from
5299 .if \to-\from
5300 sum "(\from+1)",\to
5301 .endif
5302 .endm
5303@end example
5304
5305@noindent
5306With that definition, @samp{SUM 0,5} is equivalent to this assembly input:
5307
5308@example
5309 .long 0
5310 .long 1
5311 .long 2
5312 .long 3
5313 .long 4
5314 .long 5
5315@end example
5316
5317@ftable @code
5318@item .macro @var{macname}
5319@itemx .macro @var{macname} @var{macargs} @dots{}
5320@cindex @code{macro} directive
5321Begin the definition of a macro called @var{macname}. If your macro
5322definition requires arguments, specify their names after the macro name,
6eaeac8a
JB
5323separated by commas or spaces. You can qualify the macro argument to
5324indicate whether all invocations must specify a non-blank value (through
5325@samp{:@code{req}}), or whether it takes all of the remaining arguments
5326(through @samp{:@code{vararg}}). You can supply a default value for any
fffeaa5f
JB
5327macro argument by following the name with @samp{=@var{deflt}}. You
5328cannot define two macros with the same @var{macname} unless it has been
96e9638b 5329subject to the @code{.purgem} directive (@pxref{Purgem}) between the two
fffeaa5f 5330definitions. For example, these are all valid @code{.macro} statements:
252b5132
RH
5331
5332@table @code
5333@item .macro comm
5334Begin the definition of a macro called @code{comm}, which takes no
5335arguments.
5336
6258339f 5337@item .macro plus1 p, p1
252b5132
RH
5338@itemx .macro plus1 p p1
5339Either statement begins the definition of a macro called @code{plus1},
5340which takes two arguments; within the macro definition, write
5341@samp{\p} or @samp{\p1} to evaluate the arguments.
5342
5343@item .macro reserve_str p1=0 p2
5344Begin the definition of a macro called @code{reserve_str}, with two
5345arguments. The first argument has a default value, but not the second.
5346After the definition is complete, you can call the macro either as
5347@samp{reserve_str @var{a},@var{b}} (with @samp{\p1} evaluating to
5348@var{a} and @samp{\p2} evaluating to @var{b}), or as @samp{reserve_str
5349,@var{b}} (with @samp{\p1} evaluating as the default, in this case
5350@samp{0}, and @samp{\p2} evaluating to @var{b}).
252b5132 5351
6eaeac8a
JB
5352@item .macro m p1:req, p2=0, p3:vararg
5353Begin the definition of a macro called @code{m}, with at least three
5354arguments. The first argument must always have a value specified, but
5355not the second, which instead has a default value. The third formal
5356will get assigned all remaining arguments specified at invocation time.
5357
252b5132
RH
5358When you call a macro, you can specify the argument values either by
5359position, or by keyword. For example, @samp{sum 9,17} is equivalent to
5360@samp{sum to=17, from=9}.
5361
6258339f
NC
5362@end table
5363
5e75c3ab
JB
5364Note that since each of the @var{macargs} can be an identifier exactly
5365as any other one permitted by the target architecture, there may be
5366occasional problems if the target hand-crafts special meanings to certain
6258339f 5367characters when they occur in a special position. For example, if the colon
5e75c3ab 5368(@code{:}) is generally permitted to be part of a symbol name, but the
6258339f 5369architecture specific code special-cases it when occurring as the final
5e75c3ab
JB
5370character of a symbol (to denote a label), then the macro parameter
5371replacement code will have no way of knowing that and consider the whole
5372construct (including the colon) an identifier, and check only this
6258339f
NC
5373identifier for being the subject to parameter substitution. So for example
5374this macro definition:
5375
5376@example
5377 .macro label l
5378\l:
5379 .endm
5380@end example
5381
5382might not work as expected. Invoking @samp{label foo} might not create a label
5383called @samp{foo} but instead just insert the text @samp{\l:} into the
5384assembler source, probably generating an error about an unrecognised
5385identifier.
5386
5387Similarly problems might occur with the period character (@samp{.})
5388which is often allowed inside opcode names (and hence identifier names). So
5389for example constructing a macro to build an opcode from a base name and a
5390length specifier like this:
5391
5392@example
5393 .macro opcode base length
5394 \base.\length
5395 .endm
5396@end example
5397
5398and invoking it as @samp{opcode store l} will not create a @samp{store.l}
5399instruction but instead generate some kind of error as the assembler tries to
5400interpret the text @samp{\base.\length}.
5401
5402There are several possible ways around this problem:
5403
5404@table @code
5405@item Insert white space
5406If it is possible to use white space characters then this is the simplest
5407solution. eg:
5408
5409@example
5410 .macro label l
5411\l :
5412 .endm
5413@end example
5414
5415@item Use @samp{\()}
5416The string @samp{\()} can be used to separate the end of a macro argument from
5417the following text. eg:
5418
5419@example
5420 .macro opcode base length
5421 \base\().\length
5422 .endm
5423@end example
5424
5425@item Use the alternate macro syntax mode
5426In the alternative macro syntax mode the ampersand character (@samp{&}) can be
5427used as a separator. eg:
5e75c3ab
JB
5428
5429@example
5430 .altmacro
5431 .macro label l
5432l&:
5433 .endm
5434@end example
6258339f 5435@end table
5e75c3ab 5436
96e9638b 5437Note: this problem of correctly identifying string parameters to pseudo ops
01642c12 5438also applies to the identifiers used in @code{.irp} (@pxref{Irp})
96e9638b 5439and @code{.irpc} (@pxref{Irpc}) as well.
5e75c3ab 5440
252b5132
RH
5441@item .endm
5442@cindex @code{endm} directive
5443Mark the end of a macro definition.
5444
5445@item .exitm
5446@cindex @code{exitm} directive
5447Exit early from the current macro definition.
5448
5449@cindex number of macros executed
5450@cindex macros, count executed
5451@item \@@
a4fb0134 5452@command{@value{AS}} maintains a counter of how many macros it has
252b5132
RH
5453executed in this pseudo-variable; you can copy that number to your
5454output with @samp{\@@}, but @emph{only within a macro definition}.
5455
252b5132
RH
5456@item LOCAL @var{name} [ , @dots{} ]
5457@emph{Warning: @code{LOCAL} is only available if you select ``alternate
caa32fe5
NC
5458macro syntax'' with @samp{--alternate} or @code{.altmacro}.}
5459@xref{Altmacro,,@code{.altmacro}}.
5460@end ftable
252b5132 5461
ccf8a69b
BW
5462@node MRI
5463@section @code{.mri @var{val}}
caa32fe5 5464
ccf8a69b
BW
5465@cindex @code{mri} directive
5466@cindex MRI mode, temporarily
5467If @var{val} is non-zero, this tells @command{@value{AS}} to enter MRI mode. If
5468@var{val} is zero, this tells @command{@value{AS}} to exit MRI mode. This change
5469affects code assembled until the next @code{.mri} directive, or until the end
5470of the file. @xref{M, MRI mode, MRI mode}.
252b5132 5471
caa32fe5
NC
5472@node Noaltmacro
5473@section @code{.noaltmacro}
96e9638b 5474Disable alternate macro mode. @xref{Altmacro}.
caa32fe5 5475
252b5132
RH
5476@node Nolist
5477@section @code{.nolist}
5478
5479@cindex @code{nolist} directive
5480@cindex listing control, turning off
5481Control (in conjunction with the @code{.list} directive) whether or
5482not assembly listings are generated. These two directives maintain an
5483internal counter (which is zero initially). @code{.list} increments the
5484counter, and @code{.nolist} decrements it. Assembly listings are
5485generated whenever the counter is greater than zero.
5486
5487@node Octa
5488@section @code{.octa @var{bignums}}
5489
5490@c FIXME: double size emitted for "octa" on i960, others? Or warn?
5491@cindex @code{octa} directive
5492@cindex integer, 16-byte
5493@cindex sixteen byte integer
5494This directive expects zero or more bignums, separated by commas. For each
5495bignum, it emits a 16-byte integer.
5496
5497The term ``octa'' comes from contexts in which a ``word'' is two bytes;
5498hence @emph{octa}-word for 16 bytes.
5499
5500@node Org
5501@section @code{.org @var{new-lc} , @var{fill}}
5502
5503@cindex @code{org} directive
5504@cindex location counter, advancing
5505@cindex advancing location counter
5506@cindex current address, advancing
5507Advance the location counter of the current section to
5508@var{new-lc}. @var{new-lc} is either an absolute expression or an
5509expression with the same section as the current subsection. That is,
5510you can't use @code{.org} to cross sections: if @var{new-lc} has the
5511wrong section, the @code{.org} directive is ignored. To be compatible
5512with former assemblers, if the section of @var{new-lc} is absolute,
a4fb0134 5513@command{@value{AS}} issues a warning, then pretends the section of @var{new-lc}
252b5132
RH
5514is the same as the current subsection.
5515
5516@code{.org} may only increase the location counter, or leave it
5517unchanged; you cannot use @code{.org} to move the location counter
5518backwards.
5519
5520@c double negative used below "not undefined" because this is a specific
5521@c reference to "undefined" (as SEG_UNKNOWN is called in this manual)
5522@c section. doc@cygnus.com 18feb91
a4fb0134 5523Because @command{@value{AS}} tries to assemble programs in one pass, @var{new-lc}
252b5132
RH
5524may not be undefined. If you really detest this restriction we eagerly await
5525a chance to share your improved assembler.
5526
5527Beware that the origin is relative to the start of the section, not
5528to the start of the subsection. This is compatible with other
5529people's assemblers.
5530
5531When the location counter (of the current subsection) is advanced, the
5532intervening bytes are filled with @var{fill} which should be an
5533absolute expression. If the comma and @var{fill} are omitted,
5534@var{fill} defaults to zero.
5535
5536@node P2align
5537@section @code{.p2align[wl] @var{abs-expr}, @var{abs-expr}, @var{abs-expr}}
5538
5539@cindex padding the location counter given a power of two
5540@cindex @code{p2align} directive
5541Pad the location counter (in the current subsection) to a particular
5542storage boundary. The first expression (which must be absolute) is the
5543number of low-order zero bits the location counter must have after
5544advancement. For example @samp{.p2align 3} advances the location
5545counter until it a multiple of 8. If the location counter is already a
5546multiple of 8, no change is needed.
5547
5548The second expression (also absolute) gives the fill value to be stored in the
5549padding bytes. It (and the comma) may be omitted. If it is omitted, the
5550padding bytes are normally zero. However, on some systems, if the section is
5551marked as containing code and the fill value is omitted, the space is filled
5552with no-op instructions.
5553
5554The third expression is also absolute, and is also optional. If it is present,
5555it is the maximum number of bytes that should be skipped by this alignment
5556directive. If doing the alignment would require skipping more bytes than the
5557specified maximum, then the alignment is not done at all. You can omit the
5558fill value (the second argument) entirely by simply using two commas after the
5559required alignment; this can be useful if you want the alignment to be filled
5560with no-op instructions when appropriate.
5561
5562@cindex @code{p2alignw} directive
5563@cindex @code{p2alignl} directive
5564The @code{.p2alignw} and @code{.p2alignl} directives are variants of the
5565@code{.p2align} directive. The @code{.p2alignw} directive treats the fill
5566pattern as a two byte word value. The @code{.p2alignl} directives treats the
5567fill pattern as a four byte longword value. For example, @code{.p2alignw
55682,0x368d} will align to a multiple of 4. If it skips two bytes, they will be
5569filled in with the value 0x368d (the exact placement of the bytes depends upon
5570the endianness of the processor). If it skips 1 or 3 bytes, the fill value is
5571undefined.
5572
ccf8a69b
BW
5573@ifset ELF
5574@node PopSection
5575@section @code{.popsection}
5576
5577@cindex @code{popsection} directive
5578@cindex Section Stack
5579This is one of the ELF section stack manipulation directives. The others are
01642c12
RM
5580@code{.section} (@pxref{Section}), @code{.subsection} (@pxref{SubSection}),
5581@code{.pushsection} (@pxref{PushSection}), and @code{.previous}
ccf8a69b
BW
5582(@pxref{Previous}).
5583
5584This directive replaces the current section (and subsection) with the top
5585section (and subsection) on the section stack. This section is popped off the
01642c12 5586stack.
ccf8a69b
BW
5587@end ifset
5588
c91d2e08
NC
5589@ifset ELF
5590@node Previous
5591@section @code{.previous}
5592
c1253627 5593@cindex @code{previous} directive
c91d2e08
NC
5594@cindex Section Stack
5595This is one of the ELF section stack manipulation directives. The others are
a349d9dd
PB
5596@code{.section} (@pxref{Section}), @code{.subsection} (@pxref{SubSection}),
5597@code{.pushsection} (@pxref{PushSection}), and @code{.popsection}
5598(@pxref{PopSection}).
c91d2e08
NC
5599
5600This directive swaps the current section (and subsection) with most recently
8b040e0a 5601referenced section/subsection pair prior to this one. Multiple
c91d2e08 5602@code{.previous} directives in a row will flip between two sections (and their
8b040e0a
NC
5603subsections). For example:
5604
5605@smallexample
5606.section A
5607 .subsection 1
5608 .word 0x1234
5609 .subsection 2
5610 .word 0x5678
5611.previous
5612 .word 0x9abc
5613@end smallexample
5614
5615Will place 0x1234 and 0x9abc into subsection 1 and 0x5678 into subsection 2 of
5616section A. Whilst:
5617
5618@smallexample
5619.section A
5620.subsection 1
5621 # Now in section A subsection 1
5622 .word 0x1234
5623.section B
5624.subsection 0
5625 # Now in section B subsection 0
5626 .word 0x5678
5627.subsection 1
5628 # Now in section B subsection 1
5629 .word 0x9abc
5630.previous
5631 # Now in section B subsection 0
5632 .word 0xdef0
5633@end smallexample
5634
5635Will place 0x1234 into section A, 0x5678 and 0xdef0 into subsection 0 of
5636section B and 0x9abc into subsection 1 of section B.
c91d2e08
NC
5637
5638In terms of the section stack, this directive swaps the current section with
5639the top section on the section stack.
5640@end ifset
5641
252b5132
RH
5642@node Print
5643@section @code{.print @var{string}}
5644
5645@cindex @code{print} directive
a4fb0134 5646@command{@value{AS}} will print @var{string} on the standard output during
252b5132
RH
5647assembly. You must put @var{string} in double quotes.
5648
c91d2e08
NC
5649@ifset ELF
5650@node Protected
5651@section @code{.protected @var{names}}
5652
c1253627
NC
5653@cindex @code{protected} directive
5654@cindex visibility
ed9589d4 5655This is one of the ELF visibility directives. The other two are
a349d9dd 5656@code{.hidden} (@pxref{Hidden}) and @code{.internal} (@pxref{Internal}).
c91d2e08
NC
5657
5658This directive overrides the named symbols default visibility (which is set by
5659their binding: local, global or weak). The directive sets the visibility to
5660@code{protected} which means that any references to the symbols from within the
5661components that defines them must be resolved to the definition in that
5662component, even if a definition in another component would normally preempt
01642c12 5663this.
c91d2e08
NC
5664@end ifset
5665
252b5132
RH
5666@node Psize
5667@section @code{.psize @var{lines} , @var{columns}}
5668
5669@cindex @code{psize} directive
5670@cindex listing control: paper size
5671@cindex paper size, for listings
5672Use this directive to declare the number of lines---and, optionally, the
5673number of columns---to use for each page, when generating listings.
5674
5675If you do not use @code{.psize}, listings use a default line-count
5676of 60. You may omit the comma and @var{columns} specification; the
5677default width is 200 columns.
5678
a4fb0134 5679@command{@value{AS}} generates formfeeds whenever the specified number of
252b5132
RH
5680lines is exceeded (or whenever you explicitly request one, using
5681@code{.eject}).
5682
5683If you specify @var{lines} as @code{0}, no formfeeds are generated save
5684those explicitly specified with @code{.eject}.
5685
5686@node Purgem
5687@section @code{.purgem @var{name}}
5688
5689@cindex @code{purgem} directive
5690Undefine the macro @var{name}, so that later uses of the string will not be
5691expanded. @xref{Macro}.
5692
c91d2e08
NC
5693@ifset ELF
5694@node PushSection
9cfc3331 5695@section @code{.pushsection @var{name} [, @var{subsection}] [, "@var{flags}"[, @@@var{type}[,@var{arguments}]]]}
c91d2e08 5696
c1253627 5697@cindex @code{pushsection} directive
c91d2e08
NC
5698@cindex Section Stack
5699This is one of the ELF section stack manipulation directives. The others are
01642c12
RM
5700@code{.section} (@pxref{Section}), @code{.subsection} (@pxref{SubSection}),
5701@code{.popsection} (@pxref{PopSection}), and @code{.previous}
a349d9dd 5702(@pxref{Previous}).
c91d2e08 5703
e9863d7f
DJ
5704This directive pushes the current section (and subsection) onto the
5705top of the section stack, and then replaces the current section and
9cfc3331
L
5706subsection with @code{name} and @code{subsection}. The optional
5707@code{flags}, @code{type} and @code{arguments} are treated the same
5708as in the @code{.section} (@pxref{Section}) directive.
c91d2e08
NC
5709@end ifset
5710
252b5132
RH
5711@node Quad
5712@section @code{.quad @var{bignums}}
5713
5714@cindex @code{quad} directive
5715@code{.quad} expects zero or more bignums, separated by commas. For
5716each bignum, it emits
5717@ifclear bignum-16
5718an 8-byte integer. If the bignum won't fit in 8 bytes, it prints a
5719warning message; and just takes the lowest order 8 bytes of the bignum.
5720@cindex eight-byte integer
5721@cindex integer, 8-byte
5722
5723The term ``quad'' comes from contexts in which a ``word'' is two bytes;
5724hence @emph{quad}-word for 8 bytes.
5725@end ifclear
5726@ifset bignum-16
5727a 16-byte integer. If the bignum won't fit in 16 bytes, it prints a
5728warning message; and just takes the lowest order 16 bytes of the bignum.
5729@cindex sixteen-byte integer
5730@cindex integer, 16-byte
5731@end ifset
5732
05e9452c
AM
5733@node Reloc
5734@section @code{.reloc @var{offset}, @var{reloc_name}[, @var{expression}]}
5735
5736@cindex @code{reloc} directive
5737Generate a relocation at @var{offset} of type @var{reloc_name} with value
5738@var{expression}. If @var{offset} is a number, the relocation is generated in
5739the current section. If @var{offset} is an expression that resolves to a
5740symbol plus offset, the relocation is generated in the given symbol's section.
5741@var{expression}, if present, must resolve to a symbol plus addend or to an
5742absolute value, but note that not all targets support an addend. e.g. ELF REL
5743targets such as i386 store an addend in the section contents rather than in the
5744relocation. This low level interface does not support addends stored in the
5745section.
5746
252b5132
RH
5747@node Rept
5748@section @code{.rept @var{count}}
5749
5750@cindex @code{rept} directive
5751Repeat the sequence of lines between the @code{.rept} directive and the next
5752@code{.endr} directive @var{count} times.
5753
5754For example, assembling
5755
5756@example
5757 .rept 3
5758 .long 0
5759 .endr
5760@end example
5761
5762is equivalent to assembling
5763
5764@example
5765 .long 0
5766 .long 0
5767 .long 0
5768@end example
5769
5770@node Sbttl
5771@section @code{.sbttl "@var{subheading}"}
5772
5773@cindex @code{sbttl} directive
5774@cindex subtitles for listings
5775@cindex listing control: subtitle
5776Use @var{subheading} as the title (third line, immediately after the
5777title line) when generating assembly listings.
5778
5779This directive affects subsequent pages, as well as the current page if
5780it appears within ten lines of the top of a page.
5781
5782@ifset COFF
5783@node Scl
5784@section @code{.scl @var{class}}
5785
5786@cindex @code{scl} directive
5787@cindex symbol storage class (COFF)
5788@cindex COFF symbol storage class
5789Set the storage-class value for a symbol. This directive may only be
5790used inside a @code{.def}/@code{.endef} pair. Storage class may flag
5791whether a symbol is static or external, or it may record further
5792symbolic debugging information.
5793@ifset BOUT
5794
5795The @samp{.scl} directive is primarily associated with COFF output; when
a4fb0134 5796configured to generate @code{b.out} output format, @command{@value{AS}}
252b5132
RH
5797accepts this directive but ignores it.
5798@end ifset
5799@end ifset
5800
c1253627 5801@ifset COFF-ELF
252b5132 5802@node Section
c1253627 5803@section @code{.section @var{name}}
252b5132 5804
252b5132
RH
5805@cindex named section
5806Use the @code{.section} directive to assemble the following code into a section
5807named @var{name}.
5808
5809This directive is only supported for targets that actually support arbitrarily
5810named sections; on @code{a.out} targets, for example, it is not accepted, even
5811with a standard @code{a.out} section name.
5812
c1253627
NC
5813@ifset COFF
5814@ifset ELF
5815@c only print the extra heading if both COFF and ELF are set
5816@subheading COFF Version
5817@end ifset
5818
5819@cindex @code{section} directive (COFF version)
252b5132
RH
5820For COFF targets, the @code{.section} directive is used in one of the following
5821ways:
c91d2e08 5822
252b5132
RH
5823@smallexample
5824.section @var{name}[, "@var{flags}"]
4e188d17 5825.section @var{name}[, @var{subsection}]
252b5132
RH
5826@end smallexample
5827
5828If the optional argument is quoted, it is taken as flags to use for the
5829section. Each flag is a single character. The following flags are recognized:
5830@table @code
5831@item b
5832bss section (uninitialized data)
5833@item n
5834section is not loaded
5835@item w
5836writable section
5837@item d
5838data section
5839@item r
5840read-only section
5841@item x
5842executable section
2dcc60be
ILT
5843@item s
5844shared section (meaningful for PE targets)
6ff96af6
NC
5845@item a
5846ignored. (For compatibility with the ELF version)
63ad59ae
KT
5847@item y
5848section is not readable (meaningful for PE targets)
31907d5e
DK
5849@item 0-9
5850single-digit power-of-two section alignment (GNU extension)
252b5132
RH
5851@end table
5852
5853If no flags are specified, the default flags depend upon the section name. If
5854the section name is not recognized, the default will be for the section to be
7e84d676
NC
5855loaded and writable. Note the @code{n} and @code{w} flags remove attributes
5856from the section, rather than adding them, so if they are used on their own it
5857will be as if no flags had been specified at all.
252b5132
RH
5858
5859If the optional argument to the @code{.section} directive is not quoted, it is
4e188d17 5860taken as a subsection number (@pxref{Sub-Sections}).
c1253627 5861@end ifset
252b5132
RH
5862
5863@ifset ELF
c1253627
NC
5864@ifset COFF
5865@c only print the extra heading if both COFF and ELF are set
5866@subheading ELF Version
5867@end ifset
5868
c91d2e08
NC
5869@cindex Section Stack
5870This is one of the ELF section stack manipulation directives. The others are
01642c12 5871@code{.subsection} (@pxref{SubSection}), @code{.pushsection}
a349d9dd
PB
5872(@pxref{PushSection}), @code{.popsection} (@pxref{PopSection}), and
5873@code{.previous} (@pxref{Previous}).
c91d2e08 5874
c1253627 5875@cindex @code{section} directive (ELF version)
252b5132 5876For ELF targets, the @code{.section} directive is used like this:
c91d2e08 5877
252b5132 5878@smallexample
7047dd1e 5879.section @var{name} [, "@var{flags}"[, @@@var{type}[,@var{flag_specific_arguments}]]]
252b5132 5880@end smallexample
c91d2e08 5881
252b5132 5882The optional @var{flags} argument is a quoted string which may contain any
a349d9dd 5883combination of the following characters:
252b5132
RH
5884@table @code
5885@item a
5886section is allocatable
18ae9cc1
L
5887@item e
5888section is excluded from executable and shared library.
252b5132
RH
5889@item w
5890section is writable
5891@item x
5892section is executable
ec38dd05
JJ
5893@item M
5894section is mergeable
5895@item S
5896section contains zero terminated strings
22fe14ad
NC
5897@item G
5898section is a member of a section group
5899@item T
5900section is used for thread-local-storage
01642c12
RM
5901@item ?
5902section is a member of the previously-current section's group, if any
252b5132
RH
5903@end table
5904
5905The optional @var{type} argument may contain one of the following constants:
5906@table @code
5907@item @@progbits
5908section contains data
5909@item @@nobits
5910section does not contain data (i.e., section only occupies space)
22fe14ad
NC
5911@item @@note
5912section contains data which is used by things other than the program
10b016c2
PB
5913@item @@init_array
5914section contains an array of pointers to init functions
5915@item @@fini_array
5916section contains an array of pointers to finish functions
5917@item @@preinit_array
5918section contains an array of pointers to pre-init functions
252b5132
RH
5919@end table
5920
10b016c2
PB
5921Many targets only support the first three section types.
5922
ececec60
NC
5923Note on targets where the @code{@@} character is the start of a comment (eg
5924ARM) then another character is used instead. For example the ARM port uses the
5925@code{%} character.
5926
22fe14ad 5927If @var{flags} contains the @code{M} symbol then the @var{type} argument must
96e9638b 5928be specified as well as an extra argument---@var{entsize}---like this:
22fe14ad
NC
5929
5930@smallexample
5931.section @var{name} , "@var{flags}"M, @@@var{type}, @var{entsize}
5932@end smallexample
5933
5934Sections with the @code{M} flag but not @code{S} flag must contain fixed size
5935constants, each @var{entsize} octets long. Sections with both @code{M} and
5936@code{S} must contain zero terminated strings where each character is
5937@var{entsize} bytes long. The linker may remove duplicates within sections with
5938the same name, same entity size and same flags. @var{entsize} must be an
90dce00a
AM
5939absolute expression. For sections with both @code{M} and @code{S}, a string
5940which is a suffix of a larger string is considered a duplicate. Thus
5941@code{"def"} will be merged with @code{"abcdef"}; A reference to the first
5942@code{"def"} will be changed to a reference to @code{"abcdef"+3}.
22fe14ad
NC
5943
5944If @var{flags} contains the @code{G} symbol then the @var{type} argument must
5945be present along with an additional field like this:
5946
5947@smallexample
5948.section @var{name} , "@var{flags}"G, @@@var{type}, @var{GroupName}[, @var{linkage}]
5949@end smallexample
5950
5951The @var{GroupName} field specifies the name of the section group to which this
5952particular section belongs. The optional linkage field can contain:
5953@table @code
5954@item comdat
5955indicates that only one copy of this section should be retained
5956@item .gnu.linkonce
5957an alias for comdat
5958@end table
5959
96e9638b 5960Note: if both the @var{M} and @var{G} flags are present then the fields for
22fe14ad
NC
5961the Merge flag should come first, like this:
5962
5963@smallexample
5964.section @var{name} , "@var{flags}"MG, @@@var{type}, @var{entsize}, @var{GroupName}[, @var{linkage}]
5965@end smallexample
ec38dd05 5966
01642c12
RM
5967If @var{flags} contains the @code{?} symbol then it may not also contain the
5968@code{G} symbol and the @var{GroupName} or @var{linkage} fields should not be
5969present. Instead, @code{?} says to consider the section that's current before
5970this directive. If that section used @code{G}, then the new section will use
5971@code{G} with those same @var{GroupName} and @var{linkage} fields implicitly.
5972If not, then the @code{?} symbol has no effect.
5973
252b5132
RH
5974If no flags are specified, the default flags depend upon the section name. If
5975the section name is not recognized, the default will be for the section to have
5976none of the above flags: it will not be allocated in memory, nor writable, nor
5977executable. The section will contain data.
5978
5979For ELF targets, the assembler supports another type of @code{.section}
5980directive for compatibility with the Solaris assembler:
c91d2e08 5981
252b5132
RH
5982@smallexample
5983.section "@var{name}"[, @var{flags}...]
5984@end smallexample
c91d2e08 5985
252b5132
RH
5986Note that the section name is quoted. There may be a sequence of comma
5987separated flags:
5988@table @code
5989@item #alloc
5990section is allocatable
5991@item #write
5992section is writable
5993@item #execinstr
5994section is executable
18ae9cc1
L
5995@item #exclude
5996section is excluded from executable and shared library.
22fe14ad
NC
5997@item #tls
5998section is used for thread local storage
252b5132 5999@end table
c91d2e08 6000
e9863d7f
DJ
6001This directive replaces the current section and subsection. See the
6002contents of the gas testsuite directory @code{gas/testsuite/gas/elf} for
6003some examples of how this directive and the other section stack directives
6004work.
c1253627
NC
6005@end ifset
6006@end ifset
252b5132
RH
6007
6008@node Set
6009@section @code{.set @var{symbol}, @var{expression}}
6010
6011@cindex @code{set} directive
6012@cindex symbol value, setting
6013Set the value of @var{symbol} to @var{expression}. This
6014changes @var{symbol}'s value and type to conform to
6015@var{expression}. If @var{symbol} was flagged as external, it remains
6016flagged (@pxref{Symbol Attributes}).
6017
6018You may @code{.set} a symbol many times in the same assembly.
6019
6020If you @code{.set} a global symbol, the value stored in the object
6021file is the last value stored into it.
6022
3c9b82ba
NC
6023@ifset Z80
6024On Z80 @code{set} is a real instruction, use
6025@samp{@var{symbol} defl @var{expression}} instead.
6026@end ifset
6027
252b5132
RH
6028@node Short
6029@section @code{.short @var{expressions}}
6030
6031@cindex @code{short} directive
6032@ifset GENERIC
6033@code{.short} is normally the same as @samp{.word}.
6034@xref{Word,,@code{.word}}.
6035
6036In some configurations, however, @code{.short} and @code{.word} generate
96e9638b 6037numbers of different lengths. @xref{Machine Dependencies}.
252b5132
RH
6038@end ifset
6039@ifclear GENERIC
6040@ifset W16
6041@code{.short} is the same as @samp{.word}. @xref{Word,,@code{.word}}.
6042@end ifset
6043@ifset W32
6044This expects zero or more @var{expressions}, and emits
6045a 16 bit number for each.
6046@end ifset
6047@end ifclear
6048
6049@node Single
6050@section @code{.single @var{flonums}}
6051
6052@cindex @code{single} directive
6053@cindex floating point numbers (single)
6054This directive assembles zero or more flonums, separated by commas. It
6055has the same effect as @code{.float}.
6056@ifset GENERIC
6057The exact kind of floating point numbers emitted depends on how
a4fb0134 6058@command{@value{AS}} is configured. @xref{Machine Dependencies}.
252b5132
RH
6059@end ifset
6060@ifclear GENERIC
6061@ifset IEEEFLOAT
6062On the @value{TARGET} family, @code{.single} emits 32-bit floating point
6063numbers in @sc{ieee} format.
6064@end ifset
6065@end ifclear
6066
c1253627 6067@ifset COFF-ELF
252b5132 6068@node Size
c1253627 6069@section @code{.size}
c91d2e08 6070
c1253627
NC
6071This directive is used to set the size associated with a symbol.
6072
6073@ifset COFF
6074@ifset ELF
6075@c only print the extra heading if both COFF and ELF are set
6076@subheading COFF Version
6077@end ifset
6078
6079@cindex @code{size} directive (COFF version)
6080For COFF targets, the @code{.size} directive is only permitted inside
6081@code{.def}/@code{.endef} pairs. It is used like this:
6082
6083@smallexample
6084.size @var{expression}
6085@end smallexample
252b5132 6086
c91d2e08 6087@ifset BOUT
252b5132 6088@samp{.size} is only meaningful when generating COFF format output; when
a4fb0134 6089@command{@value{AS}} is generating @code{b.out}, it accepts this directive but
252b5132
RH
6090ignores it.
6091@end ifset
c1253627 6092@end ifset
c91d2e08 6093
c1253627
NC
6094@ifset ELF
6095@ifset COFF
6096@c only print the extra heading if both COFF and ELF are set
6097@subheading ELF Version
6098@end ifset
6099
6100@cindex @code{size} directive (ELF version)
6101For ELF targets, the @code{.size} directive is used like this:
c91d2e08 6102
c1253627
NC
6103@smallexample
6104.size @var{name} , @var{expression}
6105@end smallexample
6106
6107This directive sets the size associated with a symbol @var{name}.
c91d2e08
NC
6108The size in bytes is computed from @var{expression} which can make use of label
6109arithmetic. This directive is typically used to set the size of function
6110symbols.
c1253627
NC
6111@end ifset
6112@end ifset
252b5132 6113
252b5132
RH
6114@ifclear no-space-dir
6115@node Skip
6116@section @code{.skip @var{size} , @var{fill}}
6117
6118@cindex @code{skip} directive
6119@cindex filling memory
6120This directive emits @var{size} bytes, each of value @var{fill}. Both
6121@var{size} and @var{fill} are absolute expressions. If the comma and
6122@var{fill} are omitted, @var{fill} is assumed to be zero. This is the same as
6123@samp{.space}.
884f0d36 6124@end ifclear
252b5132 6125
ccf8a69b
BW
6126@node Sleb128
6127@section @code{.sleb128 @var{expressions}}
6128
6129@cindex @code{sleb128} directive
01642c12 6130@var{sleb128} stands for ``signed little endian base 128.'' This is a
ccf8a69b
BW
6131compact, variable length representation of numbers used by the DWARF
6132symbolic debugging format. @xref{Uleb128, ,@code{.uleb128}}.
6133
884f0d36 6134@ifclear no-space-dir
252b5132
RH
6135@node Space
6136@section @code{.space @var{size} , @var{fill}}
6137
6138@cindex @code{space} directive
6139@cindex filling memory
6140This directive emits @var{size} bytes, each of value @var{fill}. Both
6141@var{size} and @var{fill} are absolute expressions. If the comma
6142and @var{fill} are omitted, @var{fill} is assumed to be zero. This is the same
6143as @samp{.skip}.
6144
6145@ifset HPPA
6146@quotation
6147@emph{Warning:} @code{.space} has a completely different meaning for HPPA
6148targets; use @code{.block} as a substitute. See @cite{HP9000 Series 800
6149Assembly Language Reference Manual} (HP 92432-90001) for the meaning of the
6150@code{.space} directive. @xref{HPPA Directives,,HPPA Assembler Directives},
6151for a summary.
6152@end quotation
6153@end ifset
6154@end ifclear
6155
252b5132
RH
6156@ifset have-stabs
6157@node Stab
6158@section @code{.stabd, .stabn, .stabs}
6159
6160@cindex symbolic debuggers, information for
6161@cindex @code{stab@var{x}} directives
6162There are three directives that begin @samp{.stab}.
6163All emit symbols (@pxref{Symbols}), for use by symbolic debuggers.
a4fb0134 6164The symbols are not entered in the @command{@value{AS}} hash table: they
252b5132
RH
6165cannot be referenced elsewhere in the source file.
6166Up to five fields are required:
6167
6168@table @var
6169@item string
6170This is the symbol's name. It may contain any character except
6171@samp{\000}, so is more general than ordinary symbol names. Some
6172debuggers used to code arbitrarily complex structures into symbol names
6173using this field.
6174
6175@item type
6176An absolute expression. The symbol's type is set to the low 8 bits of
6177this expression. Any bit pattern is permitted, but @code{@value{LD}}
6178and debuggers choke on silly bit patterns.
6179
6180@item other
6181An absolute expression. The symbol's ``other'' attribute is set to the
6182low 8 bits of this expression.
6183
6184@item desc
6185An absolute expression. The symbol's descriptor is set to the low 16
6186bits of this expression.
6187
6188@item value
6189An absolute expression which becomes the symbol's value.
6190@end table
6191
6192If a warning is detected while reading a @code{.stabd}, @code{.stabn},
6193or @code{.stabs} statement, the symbol has probably already been created;
6194you get a half-formed symbol in your object file. This is
6195compatible with earlier assemblers!
6196
6197@table @code
6198@cindex @code{stabd} directive
6199@item .stabd @var{type} , @var{other} , @var{desc}
6200
6201The ``name'' of the symbol generated is not even an empty string.
6202It is a null pointer, for compatibility. Older assemblers used a
6203null pointer so they didn't waste space in object files with empty
6204strings.
6205
6206The symbol's value is set to the location counter,
6207relocatably. When your program is linked, the value of this symbol
6208is the address of the location counter when the @code{.stabd} was
6209assembled.
6210
6211@cindex @code{stabn} directive
6212@item .stabn @var{type} , @var{other} , @var{desc} , @var{value}
6213The name of the symbol is set to the empty string @code{""}.
6214
6215@cindex @code{stabs} directive
6216@item .stabs @var{string} , @var{type} , @var{other} , @var{desc} , @var{value}
6217All five fields are specified.
6218@end table
6219@end ifset
6220@c end have-stabs
6221
6222@node String
38a57ae7 6223@section @code{.string} "@var{str}", @code{.string8} "@var{str}", @code{.string16}
01642c12 6224"@var{str}", @code{.string32} "@var{str}", @code{.string64} "@var{str}"
252b5132
RH
6225
6226@cindex string, copying to object file
38a57ae7
NC
6227@cindex string8, copying to object file
6228@cindex string16, copying to object file
6229@cindex string32, copying to object file
6230@cindex string64, copying to object file
252b5132 6231@cindex @code{string} directive
38a57ae7
NC
6232@cindex @code{string8} directive
6233@cindex @code{string16} directive
6234@cindex @code{string32} directive
6235@cindex @code{string64} directive
252b5132
RH
6236
6237Copy the characters in @var{str} to the object file. You may specify more than
6238one string to copy, separated by commas. Unless otherwise specified for a
6239particular machine, the assembler marks the end of each string with a 0 byte.
6240You can use any of the escape sequences described in @ref{Strings,,Strings}.
6241
01642c12 6242The variants @code{string16}, @code{string32} and @code{string64} differ from
38a57ae7
NC
6243the @code{string} pseudo opcode in that each 8-bit character from @var{str} is
6244copied and expanded to 16, 32 or 64 bits respectively. The expanded characters
6245are stored in target endianness byte order.
6246
6247Example:
6248@smallexample
6249 .string32 "BYE"
6250expands to:
6251 .string "B\0\0\0Y\0\0\0E\0\0\0" /* On little endian targets. */
6252 .string "\0\0\0B\0\0\0Y\0\0\0E" /* On big endian targets. */
6253@end smallexample
6254
6255
252b5132
RH
6256@node Struct
6257@section @code{.struct @var{expression}}
6258
6259@cindex @code{struct} directive
6260Switch to the absolute section, and set the section offset to @var{expression},
6261which must be an absolute expression. You might use this as follows:
6262@smallexample
6263 .struct 0
6264field1:
6265 .struct field1 + 4
6266field2:
6267 .struct field2 + 4
6268field3:
6269@end smallexample
6270This would define the symbol @code{field1} to have the value 0, the symbol
6271@code{field2} to have the value 4, and the symbol @code{field3} to have the
6272value 8. Assembly would be left in the absolute section, and you would need to
6273use a @code{.section} directive of some sort to change to some other section
6274before further assembly.
6275
c91d2e08
NC
6276@ifset ELF
6277@node SubSection
6278@section @code{.subsection @var{name}}
6279
c1253627 6280@cindex @code{subsection} directive
c91d2e08
NC
6281@cindex Section Stack
6282This is one of the ELF section stack manipulation directives. The others are
01642c12
RM
6283@code{.section} (@pxref{Section}), @code{.pushsection} (@pxref{PushSection}),
6284@code{.popsection} (@pxref{PopSection}), and @code{.previous}
a349d9dd 6285(@pxref{Previous}).
c91d2e08
NC
6286
6287This directive replaces the current subsection with @code{name}. The current
6288section is not changed. The replaced subsection is put onto the section stack
6289in place of the then current top of stack subsection.
c91d2e08
NC
6290@end ifset
6291
252b5132
RH
6292@ifset ELF
6293@node Symver
6294@section @code{.symver}
6295@cindex @code{symver} directive
6296@cindex symbol versioning
6297@cindex versions of symbols
6298Use the @code{.symver} directive to bind symbols to specific version nodes
6299within a source file. This is only supported on ELF platforms, and is
6300typically used when assembling files to be linked into a shared library.
6301There are cases where it may make sense to use this in objects to be bound
6302into an application itself so as to override a versioned symbol from a
6303shared library.
6304
79082ff0 6305For ELF targets, the @code{.symver} directive can be used like this:
252b5132
RH
6306@smallexample
6307.symver @var{name}, @var{name2@@nodename}
6308@end smallexample
339681c0 6309If the symbol @var{name} is defined within the file
79082ff0 6310being assembled, the @code{.symver} directive effectively creates a symbol
252b5132
RH
6311alias with the name @var{name2@@nodename}, and in fact the main reason that we
6312just don't try and create a regular alias is that the @var{@@} character isn't
6313permitted in symbol names. The @var{name2} part of the name is the actual name
6314of the symbol by which it will be externally referenced. The name @var{name}
6315itself is merely a name of convenience that is used so that it is possible to
6316have definitions for multiple versions of a function within a single source
6317file, and so that the compiler can unambiguously know which version of a
6318function is being mentioned. The @var{nodename} portion of the alias should be
6319the name of a node specified in the version script supplied to the linker when
6320building a shared library. If you are attempting to override a versioned
6321symbol from a shared library, then @var{nodename} should correspond to the
6322nodename of the symbol you are trying to override.
339681c0
L
6323
6324If the symbol @var{name} is not defined within the file being assembled, all
6325references to @var{name} will be changed to @var{name2@@nodename}. If no
6326reference to @var{name} is made, @var{name2@@nodename} will be removed from the
6327symbol table.
79082ff0
L
6328
6329Another usage of the @code{.symver} directive is:
6330@smallexample
6331.symver @var{name}, @var{name2@@@@nodename}
6332@end smallexample
6333In this case, the symbol @var{name} must exist and be defined within
a349d9dd 6334the file being assembled. It is similar to @var{name2@@nodename}. The
79082ff0
L
6335difference is @var{name2@@@@nodename} will also be used to resolve
6336references to @var{name2} by the linker.
6337
6338The third usage of the @code{.symver} directive is:
6339@smallexample
6340.symver @var{name}, @var{name2@@@@@@nodename}
6341@end smallexample
6342When @var{name} is not defined within the
6343file being assembled, it is treated as @var{name2@@nodename}. When
6344@var{name} is defined within the file being assembled, the symbol
6345name, @var{name}, will be changed to @var{name2@@@@nodename}.
252b5132
RH
6346@end ifset
6347
6348@ifset COFF
6349@node Tag
6350@section @code{.tag @var{structname}}
6351
6352@cindex COFF structure debugging
6353@cindex structure debugging, COFF
6354@cindex @code{tag} directive
6355This directive is generated by compilers to include auxiliary debugging
6356information in the symbol table. It is only permitted inside
6357@code{.def}/@code{.endef} pairs. Tags are used to link structure
6358definitions in the symbol table with instances of those structures.
6359@ifset BOUT
6360
6361@samp{.tag} is only used when generating COFF format output; when
a4fb0134 6362@command{@value{AS}} is generating @code{b.out}, it accepts this directive but
252b5132
RH
6363ignores it.
6364@end ifset
6365@end ifset
6366
6367@node Text
6368@section @code{.text @var{subsection}}
6369
6370@cindex @code{text} directive
a4fb0134 6371Tells @command{@value{AS}} to assemble the following statements onto the end of
252b5132
RH
6372the text subsection numbered @var{subsection}, which is an absolute
6373expression. If @var{subsection} is omitted, subsection number zero
6374is used.
6375
6376@node Title
6377@section @code{.title "@var{heading}"}
6378
6379@cindex @code{title} directive
6380@cindex listing control: title line
6381Use @var{heading} as the title (second line, immediately after the
6382source file name and pagenumber) when generating assembly listings.
6383
6384This directive affects subsequent pages, as well as the current page if
6385it appears within ten lines of the top of a page.
6386
c1253627 6387@ifset COFF-ELF
252b5132 6388@node Type
c1253627
NC
6389@section @code{.type}
6390
6391This directive is used to set the type of a symbol.
6392
6393@ifset COFF
6394@ifset ELF
6395@c only print the extra heading if both COFF and ELF are set
6396@subheading COFF Version
6397@end ifset
252b5132
RH
6398
6399@cindex COFF symbol type
6400@cindex symbol type, COFF
c1253627
NC
6401@cindex @code{type} directive (COFF version)
6402For COFF targets, this directive is permitted only within
6403@code{.def}/@code{.endef} pairs. It is used like this:
6404
6405@smallexample
6406.type @var{int}
6407@end smallexample
6408
6409This records the integer @var{int} as the type attribute of a symbol table
6410entry.
252b5132 6411
c91d2e08 6412@ifset BOUT
252b5132 6413@samp{.type} is associated only with COFF format output; when
a4fb0134 6414@command{@value{AS}} is configured for @code{b.out} output, it accepts this
252b5132
RH
6415directive but ignores it.
6416@end ifset
c1253627 6417@end ifset
c91d2e08 6418
c1253627
NC
6419@ifset ELF
6420@ifset COFF
6421@c only print the extra heading if both COFF and ELF are set
6422@subheading ELF Version
6423@end ifset
c91d2e08
NC
6424
6425@cindex ELF symbol type
6426@cindex symbol type, ELF
c1253627
NC
6427@cindex @code{type} directive (ELF version)
6428For ELF targets, the @code{.type} directive is used like this:
6429
6430@smallexample
6431.type @var{name} , @var{type description}
6432@end smallexample
6433
6434This sets the type of symbol @var{name} to be either a
a349d9dd 6435function symbol or an object symbol. There are five different syntaxes
c91d2e08 6436supported for the @var{type description} field, in order to provide
28c9d252 6437compatibility with various other assemblers.
58ab4f3d
MM
6438
6439Because some of the characters used in these syntaxes (such as @samp{@@} and
6440@samp{#}) are comment characters for some architectures, some of the syntaxes
6441below do not work on all architectures. The first variant will be accepted by
6442the GNU assembler on all architectures so that variant should be used for
6443maximum portability, if you do not need to assemble your code with other
6444assemblers.
6445
6446The syntaxes supported are:
c91d2e08
NC
6447
6448@smallexample
5671778d
NC
6449 .type <name> STT_<TYPE_IN_UPPER_CASE>
6450 .type <name>,#<type>
6451 .type <name>,@@<type>
e7c33416 6452 .type <name>,%<type>
5671778d
NC
6453 .type <name>,"<type>"
6454@end smallexample
6455
6456The types supported are:
58ab4f3d 6457
5671778d
NC
6458@table @gcctabopt
6459@item STT_FUNC
6460@itemx function
6461Mark the symbol as being a function name.
c91d2e08 6462
d8045f23
NC
6463@item STT_GNU_IFUNC
6464@itemx gnu_indirect_function
6465Mark the symbol as an indirect function when evaluated during reloc
6466processing. (This is only supported on Linux targeted assemblers).
6467
5671778d
NC
6468@item STT_OBJECT
6469@itemx object
6470Mark the symbol as being a data object.
6471
6472@item STT_TLS
6473@itemx tls_object
6474Mark the symbol as being a thead-local data object.
6475
6476@item STT_COMMON
6477@itemx common
6478Mark the symbol as being a common data object.
e7c33416
NC
6479
6480@item STT_NOTYPE
6481@itemx notype
6482Does not mark the symbol in any way. It is supported just for completeness.
6483
3e7a7d11
NC
6484@item gnu_unique_object
6485Marks the symbol as being a globally unique data object. The dynamic linker
6486will make sure that in the entire process there is just one symbol with this
6487name and type in use. (This is only supported on Linux targeted assemblers).
6488
5671778d
NC
6489@end table
6490
6491Note: Some targets support extra types in addition to those listed above.
c91d2e08 6492
c1253627
NC
6493@end ifset
6494@end ifset
c91d2e08
NC
6495
6496@node Uleb128
6497@section @code{.uleb128 @var{expressions}}
6498
6499@cindex @code{uleb128} directive
01642c12 6500@var{uleb128} stands for ``unsigned little endian base 128.'' This is a
c91d2e08 6501compact, variable length representation of numbers used by the DWARF
96e9638b 6502symbolic debugging format. @xref{Sleb128, ,@code{.sleb128}}.
252b5132
RH
6503
6504@ifset COFF
6505@node Val
6506@section @code{.val @var{addr}}
6507
6508@cindex @code{val} directive
6509@cindex COFF value attribute
6510@cindex value attribute, COFF
6511This directive, permitted only within @code{.def}/@code{.endef} pairs,
6512records the address @var{addr} as the value attribute of a symbol table
6513entry.
6514@ifset BOUT
6515
a4fb0134 6516@samp{.val} is used only for COFF output; when @command{@value{AS}} is
252b5132
RH
6517configured for @code{b.out}, it accepts this directive but ignores it.
6518@end ifset
6519@end ifset
6520
2e13b764 6521@ifset ELF
c91d2e08
NC
6522@node Version
6523@section @code{.version "@var{string}"}
2e13b764 6524
c1253627 6525@cindex @code{version} directive
c91d2e08
NC
6526This directive creates a @code{.note} section and places into it an ELF
6527formatted note of type NT_VERSION. The note's name is set to @code{string}.
9a297610 6528@end ifset
2e13b764 6529
c91d2e08
NC
6530@ifset ELF
6531@node VTableEntry
6532@section @code{.vtable_entry @var{table}, @var{offset}}
2e13b764 6533
653cfe85 6534@cindex @code{vtable_entry} directive
c91d2e08
NC
6535This directive finds or creates a symbol @code{table} and creates a
6536@code{VTABLE_ENTRY} relocation for it with an addend of @code{offset}.
2e13b764 6537
c91d2e08
NC
6538@node VTableInherit
6539@section @code{.vtable_inherit @var{child}, @var{parent}}
2e13b764 6540
653cfe85 6541@cindex @code{vtable_inherit} directive
c91d2e08
NC
6542This directive finds the symbol @code{child} and finds or creates the symbol
6543@code{parent} and then creates a @code{VTABLE_INHERIT} relocation for the
a349d9dd 6544parent whose addend is the value of the child symbol. As a special case the
96e9638b 6545parent name of @code{0} is treated as referring to the @code{*ABS*} section.
c91d2e08 6546@end ifset
2e13b764 6547
d190d046
HPN
6548@node Warning
6549@section @code{.warning "@var{string}"}
6550@cindex warning directive
6551Similar to the directive @code{.error}
6552(@pxref{Error,,@code{.error "@var{string}"}}), but just emits a warning.
6553
c91d2e08
NC
6554@node Weak
6555@section @code{.weak @var{names}}
2e13b764 6556
c1253627 6557@cindex @code{weak} directive
a349d9dd 6558This directive sets the weak attribute on the comma separated list of symbol
c91d2e08 6559@code{names}. If the symbols do not already exist, they will be created.
c87db184 6560
01642c12 6561On COFF targets other than PE, weak symbols are a GNU extension. This
977cdf5a 6562directive sets the weak attribute on the comma separated list of symbol
c87db184
CF
6563@code{names}. If the symbols do not already exist, they will be created.
6564
977cdf5a 6565On the PE target, weak symbols are supported natively as weak aliases.
01642c12 6566When a weak symbol is created that is not an alias, GAS creates an
977cdf5a 6567alternate symbol to hold the default value.
2e13b764 6568
06e77878
AO
6569@node Weakref
6570@section @code{.weakref @var{alias}, @var{target}}
6571
6572@cindex @code{weakref} directive
6573This directive creates an alias to the target symbol that enables the symbol to
6574be referenced with weak-symbol semantics, but without actually making it weak.
6575If direct references or definitions of the symbol are present, then the symbol
6576will not be weak, but if all references to it are through weak references, the
6577symbol will be marked as weak in the symbol table.
6578
6579The effect is equivalent to moving all references to the alias to a separate
6580assembly source file, renaming the alias to the symbol in it, declaring the
6581symbol as weak there, and running a reloadable link to merge the object files
6582resulting from the assembly of the new source file and the old source file that
6583had the references to the alias removed.
6584
6585The alias itself never makes to the symbol table, and is entirely handled
6586within the assembler.
6587
252b5132
RH
6588@node Word
6589@section @code{.word @var{expressions}}
6590
6591@cindex @code{word} directive
6592This directive expects zero or more @var{expressions}, of any section,
6593separated by commas.
6594@ifclear GENERIC
6595@ifset W32
a4fb0134 6596For each expression, @command{@value{AS}} emits a 32-bit number.
252b5132
RH
6597@end ifset
6598@ifset W16
a4fb0134 6599For each expression, @command{@value{AS}} emits a 16-bit number.
252b5132
RH
6600@end ifset
6601@end ifclear
6602@ifset GENERIC
6603
6604The size of the number emitted, and its byte order,
6605depend on what target computer the assembly is for.
6606@end ifset
6607
6608@c on amd29k, i960, sparc the "special treatment to support compilers" doesn't
6609@c happen---32-bit addressability, period; no long/short jumps.
6610@ifset DIFF-TBL-KLUGE
6611@cindex difference tables altered
6612@cindex altered difference tables
6613@quotation
6614@emph{Warning: Special Treatment to support Compilers}
6615@end quotation
6616
6617@ifset GENERIC
6618Machines with a 32-bit address space, but that do less than 32-bit
6619addressing, require the following special treatment. If the machine of
6620interest to you does 32-bit addressing (or doesn't require it;
6621@pxref{Machine Dependencies}), you can ignore this issue.
6622
6623@end ifset
6624In order to assemble compiler output into something that works,
a4fb0134 6625@command{@value{AS}} occasionally does strange things to @samp{.word} directives.
252b5132 6626Directives of the form @samp{.word sym1-sym2} are often emitted by
a4fb0134 6627compilers as part of jump tables. Therefore, when @command{@value{AS}} assembles a
252b5132 6628directive of the form @samp{.word sym1-sym2}, and the difference between
a4fb0134 6629@code{sym1} and @code{sym2} does not fit in 16 bits, @command{@value{AS}}
252b5132
RH
6630creates a @dfn{secondary jump table}, immediately before the next label.
6631This secondary jump table is preceded by a short-jump to the
6632first byte after the secondary table. This short-jump prevents the flow
6633of control from accidentally falling into the new table. Inside the
6634table is a long-jump to @code{sym2}. The original @samp{.word}
6635contains @code{sym1} minus the address of the long-jump to
6636@code{sym2}.
6637
6638If there were several occurrences of @samp{.word sym1-sym2} before the
6639secondary jump table, all of them are adjusted. If there was a
6640@samp{.word sym3-sym4}, that also did not fit in sixteen bits, a
6641long-jump to @code{sym4} is included in the secondary jump table,
6642and the @code{.word} directives are adjusted to contain @code{sym3}
6643minus the address of the long-jump to @code{sym4}; and so on, for as many
6644entries in the original jump table as necessary.
6645
6646@ifset INTERNALS
a4fb0134 6647@emph{This feature may be disabled by compiling @command{@value{AS}} with the
252b5132
RH
6648@samp{-DWORKING_DOT_WORD} option.} This feature is likely to confuse
6649assembly language programmers.
6650@end ifset
6651@end ifset
6652@c end DIFF-TBL-KLUGE
6653
6654@node Deprecated
6655@section Deprecated Directives
6656
6657@cindex deprecated directives
6658@cindex obsolescent directives
6659One day these directives won't work.
6660They are included for compatibility with older assemblers.
6661@table @t
6662@item .abort
6663@item .line
6664@end table
6665
3a99f02f
DJ
6666@ifset ELF
6667@node Object Attributes
6668@chapter Object Attributes
6669@cindex object attributes
6670
6671@command{@value{AS}} assembles source files written for a specific architecture
6672into object files for that architecture. But not all object files are alike.
6673Many architectures support incompatible variations. For instance, floating
6674point arguments might be passed in floating point registers if the object file
6675requires hardware floating point support---or floating point arguments might be
6676passed in integer registers if the object file supports processors with no
6677hardware floating point unit. Or, if two objects are built for different
6678generations of the same architecture, the combination may require the
6679newer generation at run-time.
6680
6681This information is useful during and after linking. At link time,
6682@command{@value{LD}} can warn about incompatible object files. After link
6683time, tools like @command{gdb} can use it to process the linked file
6684correctly.
6685
6686Compatibility information is recorded as a series of object attributes. Each
6687attribute has a @dfn{vendor}, @dfn{tag}, and @dfn{value}. The vendor is a
6688string, and indicates who sets the meaning of the tag. The tag is an integer,
6689and indicates what property the attribute describes. The value may be a string
6690or an integer, and indicates how the property affects this object. Missing
6691attributes are the same as attributes with a zero value or empty string value.
6692
6693Object attributes were developed as part of the ABI for the ARM Architecture.
6694The file format is documented in @cite{ELF for the ARM Architecture}.
6695
6696@menu
6697* GNU Object Attributes:: @sc{gnu} Object Attributes
6698* Defining New Object Attributes:: Defining New Object Attributes
6699@end menu
6700
6701@node GNU Object Attributes
6702@section @sc{gnu} Object Attributes
6703
6704The @code{.gnu_attribute} directive records an object attribute
6705with vendor @samp{gnu}.
6706
6707Except for @samp{Tag_compatibility}, which has both an integer and a string for
6708its value, @sc{gnu} attributes have a string value if the tag number is odd and
6709an integer value if the tag number is even. The second bit (@code{@var{tag} &
67102} is set for architecture-independent attributes and clear for
6711architecture-dependent ones.
6712
6713@subsection Common @sc{gnu} attributes
6714
6715These attributes are valid on all architectures.
6716
6717@table @r
6718@item Tag_compatibility (32)
6719The compatibility attribute takes an integer flag value and a vendor name. If
6720the flag value is 0, the file is compatible with other toolchains. If it is 1,
6721then the file is only compatible with the named toolchain. If it is greater
6722than 1, the file can only be processed by other toolchains under some private
6723arrangement indicated by the flag value and the vendor name.
6724@end table
6725
6726@subsection MIPS Attributes
6727
6728@table @r
6729@item Tag_GNU_MIPS_ABI_FP (4)
6730The floating-point ABI used by this object file. The value will be:
6731
6732@itemize @bullet
6733@item
67340 for files not affected by the floating-point ABI.
6735@item
67361 for files using the hardware floating-point with a standard double-precision
6737FPU.
6738@item
67392 for files using the hardware floating-point ABI with a single-precision FPU.
6740@item
67413 for files using the software floating-point ABI.
42554f6a
TS
6742@item
67434 for files using the hardware floating-point ABI with 64-bit wide
6744double-precision floating-point registers and 32-bit wide general
6745purpose registers.
3a99f02f
DJ
6746@end itemize
6747@end table
6748
6749@subsection PowerPC Attributes
6750
6751@table @r
6752@item Tag_GNU_Power_ABI_FP (4)
6753The floating-point ABI used by this object file. The value will be:
6754
6755@itemize @bullet
6756@item
67570 for files not affected by the floating-point ABI.
6758@item
3c7b9897 67591 for files using double-precision hardware floating-point ABI.
3a99f02f
DJ
6760@item
67612 for files using the software floating-point ABI.
3c7b9897
AM
6762@item
67633 for files using single-precision hardware floating-point ABI.
3a99f02f
DJ
6764@end itemize
6765
6766@item Tag_GNU_Power_ABI_Vector (8)
6767The vector ABI used by this object file. The value will be:
6768
6769@itemize @bullet
6770@item
67710 for files not affected by the vector ABI.
6772@item
67731 for files using general purpose registers to pass vectors.
6774@item
67752 for files using AltiVec registers to pass vectors.
6776@item
67773 for files using SPE registers to pass vectors.
6778@end itemize
6779@end table
6780
6781@node Defining New Object Attributes
6782@section Defining New Object Attributes
6783
6784If you want to define a new @sc{gnu} object attribute, here are the places you
6785will need to modify. New attributes should be discussed on the @samp{binutils}
6786mailing list.
6787
6788@itemize @bullet
6789@item
6790This manual, which is the official register of attributes.
6791@item
6792The header for your architecture @file{include/elf}, to define the tag.
6793@item
6794The @file{bfd} support file for your architecture, to merge the attribute
6795and issue any appropriate link warnings.
6796@item
6797Test cases in @file{ld/testsuite} for merging and link warnings.
6798@item
6799@file{binutils/readelf.c} to display your attribute.
6800@item
6801GCC, if you want the compiler to mark the attribute automatically.
6802@end itemize
6803
6804@end ifset
6805
252b5132
RH
6806@ifset GENERIC
6807@node Machine Dependencies
6808@chapter Machine Dependent Features
6809
6810@cindex machine dependencies
6811The machine instruction sets are (almost by definition) different on
a4fb0134
SC
6812each machine where @command{@value{AS}} runs. Floating point representations
6813vary as well, and @command{@value{AS}} often supports a few additional
252b5132
RH
6814directives or command-line options for compatibility with other
6815assemblers on a particular platform. Finally, some versions of
a4fb0134 6816@command{@value{AS}} support special pseudo-instructions for branch
252b5132
RH
6817optimization.
6818
6819This chapter discusses most of these differences, though it does not
6820include details on any machine's instruction set. For details on that
6821subject, see the hardware manufacturer's manual.
6822
6823@menu
625e1353
RH
6824@ifset ALPHA
6825* Alpha-Dependent:: Alpha Dependent Features
6826@end ifset
252b5132
RH
6827@ifset ARC
6828* ARC-Dependent:: ARC Dependent Features
6829@end ifset
6830@ifset ARM
6831* ARM-Dependent:: ARM Dependent Features
6832@end ifset
8473f7a4
DC
6833@ifset AVR
6834* AVR-Dependent:: AVR Dependent Features
6835@end ifset
3b4e1885
JZ
6836@ifset Blackfin
6837* Blackfin-Dependent:: Blackfin Dependent Features
07c1b327 6838@end ifset
3d3d428f
NC
6839@ifset CR16
6840* CR16-Dependent:: CR16 Dependent Features
6841@end ifset
8bf549a8 6842@ifset CRIS
328eb32e
HPN
6843* CRIS-Dependent:: CRIS Dependent Features
6844@end ifset
252b5132
RH
6845@ifset D10V
6846* D10V-Dependent:: D10V Dependent Features
6847@end ifset
6848@ifset D30V
6849* D30V-Dependent:: D30V Dependent Features
6850@end ifset
6851@ifset H8/300
c2dcd04e 6852* H8/300-Dependent:: Renesas H8/300 Dependent Features
252b5132 6853@end ifset
252b5132
RH
6854@ifset HPPA
6855* HPPA-Dependent:: HPPA Dependent Features
6856@end ifset
5b93d8bb
AM
6857@ifset I370
6858* ESA/390-Dependent:: IBM ESA/390 Dependent Features
6859@end ifset
252b5132 6860@ifset I80386
55b62671 6861* i386-Dependent:: Intel 80386 and AMD x86-64 Dependent Features
252b5132 6862@end ifset
e3308d0d
JE
6863@ifset I860
6864* i860-Dependent:: Intel 80860 Dependent Features
6865@end ifset
252b5132
RH
6866@ifset I960
6867* i960-Dependent:: Intel 80960 Dependent Features
6868@end ifset
5cb53c21
L
6869@ifset IA64
6870* IA-64-Dependent:: Intel IA-64 Dependent Features
6871@end ifset
a40cbfa3
NC
6872@ifset IP2K
6873* IP2K-Dependent:: IP2K Dependent Features
6874@end ifset
84e94c90
NC
6875@ifset LM32
6876* LM32-Dependent:: LM32 Dependent Features
6877@end ifset
49f58d10
JB
6878@ifset M32C
6879* M32C-Dependent:: M32C Dependent Features
6880@end ifset
ec694b89
NC
6881@ifset M32R
6882* M32R-Dependent:: M32R Dependent Features
6883@end ifset
252b5132
RH
6884@ifset M680X0
6885* M68K-Dependent:: M680x0 Dependent Features
6886@end ifset
60bcf0fa
NC
6887@ifset M68HC11
6888* M68HC11-Dependent:: M68HC11 and 68HC12 Dependent Features
6889@end ifset
7ba29e2a
NC
6890@ifset MICROBLAZE
6891* MicroBlaze-Dependent:: MICROBLAZE Dependent Features
6892@end ifset
252b5132
RH
6893@ifset MIPS
6894* MIPS-Dependent:: MIPS Dependent Features
6895@end ifset
3c3bdf30
NC
6896@ifset MMIX
6897* MMIX-Dependent:: MMIX Dependent Features
6898@end ifset
2469cfa2
NC
6899@ifset MSP430
6900* MSP430-Dependent:: MSP430 Dependent Features
6901@end ifset
252b5132 6902@ifset SH
ef230218
JR
6903* SH-Dependent:: Renesas / SuperH SH Dependent Features
6904* SH64-Dependent:: SuperH SH64 Dependent Features
252b5132 6905@end ifset
e135f41b
NC
6906@ifset PDP11
6907* PDP-11-Dependent:: PDP-11 Dependent Features
6908@end ifset
041dd5a9
ILT
6909@ifset PJ
6910* PJ-Dependent:: picoJava Dependent Features
6911@end ifset
418c1742
MG
6912@ifset PPC
6913* PPC-Dependent:: PowerPC Dependent Features
6914@end ifset
046d31c2
NC
6915@ifset RX
6916* RX-Dependent:: RX Dependent Features
6917@end ifset
11c19e16
MS
6918@ifset S390
6919* S/390-Dependent:: IBM S/390 Dependent Features
6920@end ifset
c0157db4
NC
6921@ifset SCORE
6922* SCORE-Dependent:: SCORE Dependent Features
6923@end ifset
252b5132
RH
6924@ifset SPARC
6925* Sparc-Dependent:: SPARC Dependent Features
6926@end ifset
39bec121
TW
6927@ifset TIC54X
6928* TIC54X-Dependent:: TI TMS320C54x Dependent Features
6929@end ifset
40b36596
JM
6930@ifset TIC6X
6931* TIC6X-Dependent :: TI TMS320C6x Dependent Features
6932@end ifset
252b5132
RH
6933@ifset V850
6934* V850-Dependent:: V850 Dependent Features
6935@end ifset
e0001a05
NC
6936@ifset XTENSA
6937* Xtensa-Dependent:: Xtensa Dependent Features
6938@end ifset
3c9b82ba
NC
6939@ifset Z80
6940* Z80-Dependent:: Z80 Dependent Features
6941@end ifset
252b5132
RH
6942@ifset Z8000
6943* Z8000-Dependent:: Z8000 Dependent Features
6944@end ifset
6945@ifset VAX
6946* Vax-Dependent:: VAX Dependent Features
6947@end ifset
6948@end menu
6949
6950@lowersections
6951@end ifset
6952
6953@c The following major nodes are *sections* in the GENERIC version, *chapters*
6954@c in single-cpu versions. This is mainly achieved by @lowersections. There is a
6955@c peculiarity: to preserve cross-references, there must be a node called
6956@c "Machine Dependencies". Hence the conditional nodenames in each
6957@c major node below. Node defaulting in makeinfo requires adjacency of
6958@c node and sectioning commands; hence the repetition of @chapter BLAH
6959@c in both conditional blocks.
6960
625e1353
RH
6961@ifset ALPHA
6962@include c-alpha.texi
6963@end ifset
6964
6965@ifset ARC
6966@include c-arc.texi
6967@end ifset
6968
252b5132
RH
6969@ifset ARM
6970@include c-arm.texi
6971@end ifset
6972
8473f7a4
DC
6973@ifset AVR
6974@include c-avr.texi
6975@end ifset
6976
3b4e1885 6977@ifset Blackfin
07c1b327
CM
6978@include c-bfin.texi
6979@end ifset
6980
3d3d428f
NC
6981@ifset CR16
6982@include c-cr16.texi
6983@end ifset
6984
328eb32e
HPN
6985@ifset CRIS
6986@include c-cris.texi
6987@end ifset
6988
c2dcd04e 6989@ifset Renesas-all
252b5132
RH
6990@ifclear GENERIC
6991@node Machine Dependencies
6992@chapter Machine Dependent Features
6993
c2dcd04e 6994The machine instruction sets are different on each Renesas chip family,
252b5132 6995and there are also some syntax differences among the families. This
a4fb0134 6996chapter describes the specific @command{@value{AS}} features for each
252b5132
RH
6997family.
6998
6999@menu
c2dcd04e 7000* H8/300-Dependent:: Renesas H8/300 Dependent Features
c2dcd04e 7001* SH-Dependent:: Renesas SH Dependent Features
252b5132
RH
7002@end menu
7003@lowersections
7004@end ifclear
7005@end ifset
7006
7007@ifset D10V
7008@include c-d10v.texi
7009@end ifset
7010
7011@ifset D30V
7012@include c-d30v.texi
7013@end ifset
7014
7015@ifset H8/300
7016@include c-h8300.texi
7017@end ifset
7018
252b5132
RH
7019@ifset HPPA
7020@include c-hppa.texi
7021@end ifset
7022
5b93d8bb
AM
7023@ifset I370
7024@include c-i370.texi
7025@end ifset
7026
252b5132
RH
7027@ifset I80386
7028@include c-i386.texi
7029@end ifset
7030
e3308d0d
JE
7031@ifset I860
7032@include c-i860.texi
7033@end ifset
7034
252b5132
RH
7035@ifset I960
7036@include c-i960.texi
7037@end ifset
7038
9e32ca89
NC
7039@ifset IA64
7040@include c-ia64.texi
7041@end ifset
7042
a40cbfa3
NC
7043@ifset IP2K
7044@include c-ip2k.texi
7045@end ifset
7046
84e94c90
NC
7047@ifset LM32
7048@include c-lm32.texi
7049@end ifset
7050
49f58d10
JB
7051@ifset M32C
7052@include c-m32c.texi
7053@end ifset
7054
ec694b89
NC
7055@ifset M32R
7056@include c-m32r.texi
7057@end ifset
252b5132
RH
7058
7059@ifset M680X0
7060@include c-m68k.texi
7061@end ifset
7062
60bcf0fa
NC
7063@ifset M68HC11
7064@include c-m68hc11.texi
7065@end ifset
7066
01642c12 7067@ifset MICROBLAZE
7ba29e2a
NC
7068@include c-microblaze.texi
7069@end ifset
7070
252b5132
RH
7071@ifset MIPS
7072@include c-mips.texi
7073@end ifset
7074
3c3bdf30
NC
7075@ifset MMIX
7076@include c-mmix.texi
7077@end ifset
7078
2469cfa2
NC
7079@ifset MSP430
7080@include c-msp430.texi
7081@end ifset
7082
252b5132
RH
7083@ifset NS32K
7084@include c-ns32k.texi
7085@end ifset
7086
e135f41b
NC
7087@ifset PDP11
7088@include c-pdp11.texi
7089@end ifset
7090
041dd5a9
ILT
7091@ifset PJ
7092@include c-pj.texi
7093@end ifset
7094
418c1742
MG
7095@ifset PPC
7096@include c-ppc.texi
7097@end ifset
7098
046d31c2
NC
7099@ifset RX
7100@include c-rx.texi
7101@end ifset
7102
11c19e16
MS
7103@ifset S390
7104@include c-s390.texi
7105@end ifset
7106
c0157db4
NC
7107@ifset SCORE
7108@include c-score.texi
7109@end ifset
7110
252b5132
RH
7111@ifset SH
7112@include c-sh.texi
324bfcf3 7113@include c-sh64.texi
252b5132
RH
7114@end ifset
7115
7116@ifset SPARC
7117@include c-sparc.texi
7118@end ifset
7119
39bec121
TW
7120@ifset TIC54X
7121@include c-tic54x.texi
7122@end ifset
7123
40b36596
JM
7124@ifset TIC6X
7125@include c-tic6x.texi
7126@end ifset
7127
3c9b82ba
NC
7128@ifset Z80
7129@include c-z80.texi
7130@end ifset
7131
252b5132
RH
7132@ifset Z8000
7133@include c-z8k.texi
7134@end ifset
7135
7136@ifset VAX
7137@include c-vax.texi
7138@end ifset
7139
7140@ifset V850
7141@include c-v850.texi
7142@end ifset
7143
e0001a05
NC
7144@ifset XTENSA
7145@include c-xtensa.texi
7146@end ifset
7147
252b5132
RH
7148@ifset GENERIC
7149@c reverse effect of @down at top of generic Machine-Dep chapter
7150@raisesections
7151@end ifset
7152
7153@node Reporting Bugs
7154@chapter Reporting Bugs
7155@cindex bugs in assembler
7156@cindex reporting bugs in assembler
7157
a4fb0134 7158Your bug reports play an essential role in making @command{@value{AS}} reliable.
252b5132
RH
7159
7160Reporting a bug may help you by bringing a solution to your problem, or it may
7161not. But in any case the principal function of a bug report is to help the
a4fb0134
SC
7162entire community by making the next version of @command{@value{AS}} work better.
7163Bug reports are your contribution to the maintenance of @command{@value{AS}}.
252b5132
RH
7164
7165In order for a bug report to serve its purpose, you must include the
7166information that enables us to fix the bug.
7167
7168@menu
7169* Bug Criteria:: Have you found a bug?
7170* Bug Reporting:: How to report bugs
7171@end menu
7172
7173@node Bug Criteria
c1253627 7174@section Have You Found a Bug?
252b5132
RH
7175@cindex bug criteria
7176
7177If you are not sure whether you have found a bug, here are some guidelines:
7178
7179@itemize @bullet
7180@cindex fatal signal
7181@cindex assembler crash
7182@cindex crash of assembler
7183@item
7184If the assembler gets a fatal signal, for any input whatever, that is a
a4fb0134 7185@command{@value{AS}} bug. Reliable assemblers never crash.
252b5132
RH
7186
7187@cindex error on valid input
7188@item
a4fb0134 7189If @command{@value{AS}} produces an error message for valid input, that is a bug.
252b5132
RH
7190
7191@cindex invalid input
7192@item
a4fb0134 7193If @command{@value{AS}} does not produce an error message for invalid input, that
252b5132
RH
7194is a bug. However, you should note that your idea of ``invalid input'' might
7195be our idea of ``an extension'' or ``support for traditional practice''.
7196
7197@item
7198If you are an experienced user of assemblers, your suggestions for improvement
a4fb0134 7199of @command{@value{AS}} are welcome in any case.
252b5132
RH
7200@end itemize
7201
7202@node Bug Reporting
c1253627 7203@section How to Report Bugs
252b5132
RH
7204@cindex bug reports
7205@cindex assembler bugs, reporting
7206
7207A number of companies and individuals offer support for @sc{gnu} products. If
a4fb0134 7208you obtained @command{@value{AS}} from a support organization, we recommend you
252b5132
RH
7209contact that organization first.
7210
7211You can find contact information for many support companies and
7212individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
7213distribution.
7214
ad22bfe8 7215@ifset BUGURL
a4fb0134 7216In any event, we also recommend that you send bug reports for @command{@value{AS}}
ad22bfe8
JM
7217to @value{BUGURL}.
7218@end ifset
252b5132
RH
7219
7220The fundamental principle of reporting bugs usefully is this:
7221@strong{report all the facts}. If you are not sure whether to state a
7222fact or leave it out, state it!
7223
7224Often people omit facts because they think they know what causes the problem
7225and assume that some details do not matter. Thus, you might assume that the
7226name of a symbol you use in an example does not matter. Well, probably it does
7227not, but one cannot be sure. Perhaps the bug is a stray memory reference which
7228happens to fetch from the location where that name is stored in memory;
7229perhaps, if the name were different, the contents of that location would fool
7230the assembler into doing the right thing despite the bug. Play it safe and
7231give a specific, complete example. That is the easiest thing for you to do,
7232and the most helpful.
7233
7234Keep in mind that the purpose of a bug report is to enable us to fix the bug if
7235it is new to us. Therefore, always write your bug reports on the assumption
7236that the bug has not been reported previously.
7237
7238Sometimes people give a few sketchy facts and ask, ``Does this ring a
c1253627
NC
7239bell?'' This cannot help us fix a bug, so it is basically useless. We
7240respond by asking for enough details to enable us to investigate.
7241You might as well expedite matters by sending them to begin with.
252b5132
RH
7242
7243To enable us to fix the bug, you should include all these things:
7244
7245@itemize @bullet
7246@item
a4fb0134 7247The version of @command{@value{AS}}. @command{@value{AS}} announces it if you start
252b5132
RH
7248it with the @samp{--version} argument.
7249
7250Without this, we will not know whether there is any point in looking for
a4fb0134 7251the bug in the current version of @command{@value{AS}}.
252b5132
RH
7252
7253@item
a4fb0134 7254Any patches you may have applied to the @command{@value{AS}} source.
252b5132
RH
7255
7256@item
7257The type of machine you are using, and the operating system name and
7258version number.
7259
7260@item
a4fb0134 7261What compiler (and its version) was used to compile @command{@value{AS}}---e.g.
252b5132
RH
7262``@code{gcc-2.7}''.
7263
7264@item
7265The command arguments you gave the assembler to assemble your example and
7266observe the bug. To guarantee you will not omit something important, list them
7267all. A copy of the Makefile (or the output from make) is sufficient.
7268
7269If we were to try to guess the arguments, we would probably guess wrong
7270and then we might not encounter the bug.
7271
7272@item
7273A complete input file that will reproduce the bug. If the bug is observed when
7274the assembler is invoked via a compiler, send the assembler source, not the
7275high level language source. Most compilers will produce the assembler source
7276when run with the @samp{-S} option. If you are using @code{@value{GCC}}, use
7277the options @samp{-v --save-temps}; this will save the assembler source in a
7278file with an extension of @file{.s}, and also show you exactly how
a4fb0134 7279@command{@value{AS}} is being run.
252b5132
RH
7280
7281@item
7282A description of what behavior you observe that you believe is
7283incorrect. For example, ``It gets a fatal signal.''
7284
a4fb0134 7285Of course, if the bug is that @command{@value{AS}} gets a fatal signal, then we
252b5132
RH
7286will certainly notice it. But if the bug is incorrect output, we might not
7287notice unless it is glaringly wrong. You might as well not give us a chance to
7288make a mistake.
7289
7290Even if the problem you experience is a fatal signal, you should still say so
7291explicitly. Suppose something strange is going on, such as, your copy of
b45619c0 7292@command{@value{AS}} is out of sync, or you have encountered a bug in the C
252b5132
RH
7293library on your system. (This has happened!) Your copy might crash and ours
7294would not. If you told us to expect a crash, then when ours fails to crash, we
7295would know that the bug was not happening for us. If you had not told us to
7296expect a crash, then we would not be able to draw any conclusion from our
7297observations.
7298
7299@item
a4fb0134 7300If you wish to suggest changes to the @command{@value{AS}} source, send us context
252b5132
RH
7301diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or @samp{-p}
7302option. Always send diffs from the old file to the new file. If you even
a4fb0134 7303discuss something in the @command{@value{AS}} source, refer to it by context, not
252b5132
RH
7304by line number.
7305
7306The line numbers in our development sources will not match those in your
7307sources. Your line numbers would convey no useful information to us.
7308@end itemize
7309
7310Here are some things that are not necessary:
7311
7312@itemize @bullet
7313@item
7314A description of the envelope of the bug.
7315
7316Often people who encounter a bug spend a lot of time investigating
7317which changes to the input file will make the bug go away and which
7318changes will not affect it.
7319
7320This is often time consuming and not very useful, because the way we
7321will find the bug is by running a single example under the debugger
7322with breakpoints, not by pure deduction from a series of examples.
7323We recommend that you save your time for something else.
7324
7325Of course, if you can find a simpler example to report @emph{instead}
7326of the original one, that is a convenience for us. Errors in the
7327output will be easier to spot, running under the debugger will take
7328less time, and so on.
7329
7330However, simplification is not vital; if you do not want to do this,
7331report the bug anyway and send us the entire test case you used.
7332
7333@item
7334A patch for the bug.
7335
7336A patch for the bug does help us if it is a good one. But do not omit
7337the necessary information, such as the test case, on the assumption that
7338a patch is all we need. We might see problems with your patch and decide
7339to fix the problem another way, or we might not understand it at all.
7340
a4fb0134 7341Sometimes with a program as complicated as @command{@value{AS}} it is very hard to
252b5132
RH
7342construct an example that will make the program follow a certain path through
7343the code. If you do not send us the example, we will not be able to construct
7344one, so we will not be able to verify that the bug is fixed.
7345
7346And if we cannot understand what bug you are trying to fix, or why your
7347patch should be an improvement, we will not install it. A test case will
7348help us to understand.
7349
7350@item
7351A guess about what the bug is or what it depends on.
7352
7353Such guesses are usually wrong. Even we cannot guess right about such
7354things without first using the debugger to find the facts.
7355@end itemize
7356
7357@node Acknowledgements
7358@chapter Acknowledgements
7359
653cfe85 7360If you have contributed to GAS and your name isn't listed here,
252b5132 7361it is not meant as a slight. We just don't know about it. Send mail to the
01642c12
RM
7362maintainer, and we'll correct the situation. Currently
7363@c (January 1994),
252b5132
RH
7364the maintainer is Ken Raeburn (email address @code{raeburn@@cygnus.com}).
7365
7366Dean Elsner wrote the original @sc{gnu} assembler for the VAX.@footnote{Any
7367more details?}
7368
7369Jay Fenlason maintained GAS for a while, adding support for GDB-specific debug
7370information and the 68k series machines, most of the preprocessing pass, and
7371extensive changes in @file{messages.c}, @file{input-file.c}, @file{write.c}.
7372
7373K. Richard Pixley maintained GAS for a while, adding various enhancements and
7374many bug fixes, including merging support for several processors, breaking GAS
7375up to handle multiple object file format back ends (including heavy rewrite,
7376testing, an integration of the coff and b.out back ends), adding configuration
7377including heavy testing and verification of cross assemblers and file splits
7378and renaming, converted GAS to strictly ANSI C including full prototypes, added
7379support for m680[34]0 and cpu32, did considerable work on i960 including a COFF
7380port (including considerable amounts of reverse engineering), a SPARC opcode
7381file rewrite, DECstation, rs6000, and hp300hpux host ports, updated ``know''
7382assertions and made them work, much other reorganization, cleanup, and lint.
7383
7384Ken Raeburn wrote the high-level BFD interface code to replace most of the code
7385in format-specific I/O modules.
7386
7387The original VMS support was contributed by David L. Kashtan. Eric Youngdale
7388has done much work with it since.
7389
7390The Intel 80386 machine description was written by Eliot Dresselhaus.
7391
7392Minh Tran-Le at IntelliCorp contributed some AIX 386 support.
7393
7394The Motorola 88k machine description was contributed by Devon Bowen of Buffalo
7395University and Torbjorn Granlund of the Swedish Institute of Computer Science.
7396
7397Keith Knowles at the Open Software Foundation wrote the original MIPS back end
7398(@file{tc-mips.c}, @file{tc-mips.h}), and contributed Rose format support
7399(which hasn't been merged in yet). Ralph Campbell worked with the MIPS code to
7400support a.out format.
7401
7be1c489
AM
7402Support for the Zilog Z8k and Renesas H8/300 processors (tc-z8k,
7403tc-h8300), and IEEE 695 object file format (obj-ieee), was written by
252b5132
RH
7404Steve Chamberlain of Cygnus Support. Steve also modified the COFF back end to
7405use BFD for some low-level operations, for use with the H8/300 and AMD 29k
7406targets.
7407
7408John Gilmore built the AMD 29000 support, added @code{.include} support, and
7409simplified the configuration of which versions accept which directives. He
7410updated the 68k machine description so that Motorola's opcodes always produced
c1253627 7411fixed-size instructions (e.g., @code{jsr}), while synthetic instructions
252b5132
RH
7412remained shrinkable (@code{jbsr}). John fixed many bugs, including true tested
7413cross-compilation support, and one bug in relaxation that took a week and
7414required the proverbial one-bit fix.
7415
7416Ian Lance Taylor of Cygnus Support merged the Motorola and MIT syntax for the
741768k, completed support for some COFF targets (68k, i386 SVR3, and SCO Unix),
7418added support for MIPS ECOFF and ELF targets, wrote the initial RS/6000 and
7419PowerPC assembler, and made a few other minor patches.
7420
653cfe85 7421Steve Chamberlain made GAS able to generate listings.
252b5132
RH
7422
7423Hewlett-Packard contributed support for the HP9000/300.
7424
7425Jeff Law wrote GAS and BFD support for the native HPPA object format (SOM)
7426along with a fairly extensive HPPA testsuite (for both SOM and ELF object
7427formats). This work was supported by both the Center for Software Science at
7428the University of Utah and Cygnus Support.
7429
7430Support for ELF format files has been worked on by Mark Eichin of Cygnus
7431Support (original, incomplete implementation for SPARC), Pete Hoogenboom and
7432Jeff Law at the University of Utah (HPPA mainly), Michael Meissner of the Open
7433Software Foundation (i386 mainly), and Ken Raeburn of Cygnus Support (sparc,
7434and some initial 64-bit support).
7435
c1253627 7436Linas Vepstas added GAS support for the ESA/390 ``IBM 370'' architecture.
5b93d8bb 7437
252b5132
RH
7438Richard Henderson rewrote the Alpha assembler. Klaus Kaempf wrote GAS and BFD
7439support for openVMS/Alpha.
7440
39bec121
TW
7441Timothy Wall, Michael Hayes, and Greg Smart contributed to the various tic*
7442flavors.
7443
e0001a05 7444David Heine, Sterling Augustine, Bob Wilson and John Ruttenberg from Tensilica,
b45619c0 7445Inc.@: added support for Xtensa processors.
e0001a05 7446
252b5132
RH
7447Several engineers at Cygnus Support have also provided many small bug fixes and
7448configuration enhancements.
7449
84e94c90
NC
7450Jon Beniston added support for the Lattice Mico32 architecture.
7451
252b5132
RH
7452Many others have contributed large or small bugfixes and enhancements. If
7453you have contributed significant work and are not mentioned on this list, and
7454want to be, let us know. Some of the history has been lost; we are not
7455intentionally leaving anyone out.
7456
793c5807
NC
7457@node GNU Free Documentation License
7458@appendix GNU Free Documentation License
c1253627 7459@include fdl.texi
cf055d54 7460
370b66a1
CD
7461@node AS Index
7462@unnumbered AS Index
252b5132
RH
7463
7464@printindex cp
7465
252b5132
RH
7466@bye
7467@c Local Variables:
7468@c fill-column: 79
7469@c End:
This page took 0.945731 seconds and 4 git commands to generate.