1 @c Copyright (C) 2009-2020 Free Software Foundation, Inc.
2 @c Contributed by ARM Ltd.
3 @c This is part of the GAS manual.
4 @c For copying conditions, see the file as.texinfo.
9 @node AArch64-Dependent
10 @chapter AArch64 Dependent Features
14 @node Machine Dependencies
15 @chapter AArch64 Dependent Features
18 @cindex AArch64 support
20 * AArch64 Options:: Options
21 * AArch64 Extensions:: Extensions
22 * AArch64 Syntax:: Syntax
23 * AArch64 Floating Point:: Floating Point
24 * AArch64 Directives:: AArch64 Machine Directives
25 * AArch64 Opcodes:: Opcodes
26 * AArch64 Mapping Symbols:: Mapping Symbols
31 @cindex AArch64 options (none)
32 @cindex options for AArch64 (none)
37 @cindex @option{-EB} command-line option, AArch64
39 This option specifies that the output generated by the assembler should
40 be marked as being encoded for a big-endian processor.
42 @cindex @option{-EL} command-line option, AArch64
44 This option specifies that the output generated by the assembler should
45 be marked as being encoded for a little-endian processor.
47 @cindex @option{-mabi=} command-line option, AArch64
49 Specify which ABI the source code uses. The recognized arguments
50 are: @code{ilp32} and @code{lp64}, which decides the generated object
51 file in ELF32 and ELF64 format respectively. The default is @code{lp64}.
53 @cindex @option{-mcpu=} command-line option, AArch64
54 @item -mcpu=@var{processor}[+@var{extension}@dots{}]
55 This option specifies the target processor. The assembler will issue an error
56 message if an attempt is made to assemble an instruction which will not execute
57 on the target processor. The following processor names are recognized:
86 The special name @code{all} may be used to allow the assembler to accept
87 instructions valid for any supported processor, including all optional
90 In addition to the basic instruction set, the assembler can be told to
91 accept, or restrict, various extension mnemonics that extend the
92 processor. @xref{AArch64 Extensions}.
94 If some implementations of a particular processor can have an
95 extension, then then those extensions are automatically enabled.
96 Consequently, you will not normally have to specify any additional
99 @cindex @option{-march=} command-line option, AArch64
100 @item -march=@var{architecture}[+@var{extension}@dots{}]
101 This option specifies the target architecture. The assembler will
102 issue an error message if an attempt is made to assemble an
103 instruction which will not execute on the target architecture. The
104 following architecture names are recognized: @code{armv8-a},
105 @code{armv8.1-a}, @code{armv8.2-a}, @code{armv8.3-a}, @code{armv8.4-a}
106 @code{armv8.5-a}, @code{armv8.6-a}, and @code{armv8-r}.
108 If both @option{-mcpu} and @option{-march} are specified, the
109 assembler will use the setting for @option{-mcpu}. If neither are
110 specified, the assembler will default to @option{-mcpu=all}.
112 The architecture option can be extended with the same instruction set
113 extension options as the @option{-mcpu} option. Unlike
114 @option{-mcpu}, extensions are not always enabled by default,
115 @xref{AArch64 Extensions}.
117 @cindex @code{-mverbose-error} command-line option, AArch64
118 @item -mverbose-error
119 This option enables verbose error messages for AArch64 gas. This option
120 is enabled by default.
122 @cindex @code{-mno-verbose-error} command-line option, AArch64
123 @item -mno-verbose-error
124 This option disables verbose error messages in AArch64 gas.
129 @node AArch64 Extensions
130 @section Architecture Extensions
132 The table below lists the permitted architecture extensions that are
133 supported by the assembler and the conditions under which they are
134 automatically enabled.
136 Multiple extensions may be specified, separated by a @code{+}.
137 Extension mnemonics may also be removed from those the assembler
138 accepts. This is done by prepending @code{no} to the option that adds
139 the extension. Extensions that are removed must be listed after all
140 extensions that have been added.
142 Enabling an extension that requires other extensions will
143 automatically cause those extensions to be enabled. Similarly,
144 disabling an extension that is required by other extensions will
145 automatically cause those extensions to be disabled.
147 @multitable @columnfractions .12 .17 .17 .54
148 @headitem Extension @tab Minimum Architecture @tab Enabled by default
150 @item @code{i8mm} @tab ARMv8.2-A @tab ARMv8.6-A or later
151 @tab Enable Int8 Matrix Multiply extension.
152 @item @code{f32mm} @tab ARMv8.2-A @tab No
153 @tab Enable F32 Matrix Multiply extension.
154 @item @code{f64mm} @tab ARMv8.2-A @tab No
155 @tab Enable F64 Matrix Multiply extension.
156 @item @code{bf16} @tab ARMv8.2-A @tab ARMv8.6-A or later
157 @tab Enable BFloat16 extension.
158 @item @code{compnum} @tab ARMv8.2-A @tab ARMv8.3-A or later
159 @tab Enable the complex number SIMD extensions. This implies
160 @code{fp16} and @code{simd}.
161 @item @code{crc} @tab ARMv8-A @tab ARMv8.1-A or later
162 @tab Enable CRC instructions.
163 @item @code{crypto} @tab ARMv8-A @tab No
164 @tab Enable cryptographic extensions. This implies @code{fp}, @code{simd}, @code{aes} and @code{sha2}.
165 @item @code{aes} @tab ARMv8-A @tab No
166 @tab Enable the AES cryptographic extensions. This implies @code{fp} and @code{simd}.
167 @item @code{sha2} @tab ARMv8-A @tab No
168 @tab Enable the SHA2 cryptographic extensions. This implies @code{fp} and @code{simd}.
169 @item @code{sha3} @tab ARMv8.2-A @tab No
170 @tab Enable the ARMv8.2-A SHA2 and SHA3 cryptographic extensions. This implies @code{fp}, @code{simd} and @code{sha2}.
171 @item @code{sm4} @tab ARMv8.2-A @tab No
172 @tab Enable the ARMv8.2-A SM3 and SM4 cryptographic extensions. This implies @code{fp} and @code{simd}.
173 @item @code{fp} @tab ARMv8-A @tab ARMv8-A or later
174 @tab Enable floating-point extensions.
175 @item @code{fp16} @tab ARMv8.2-A @tab ARMv8.2-A or later
176 @tab Enable ARMv8.2 16-bit floating-point support. This implies
178 @item @code{lor} @tab ARMv8-A @tab ARMv8.1-A or later
179 @tab Enable Limited Ordering Regions extensions.
180 @item @code{lse} @tab ARMv8-A @tab ARMv8.1-A or later
181 @tab Enable Large System extensions.
182 @item @code{pan} @tab ARMv8-A @tab ARMv8.1-A or later
183 @tab Enable Privileged Access Never support.
184 @item @code{profile} @tab ARMv8.2-A @tab No
185 @tab Enable statistical profiling extensions.
186 @item @code{ras} @tab ARMv8-A @tab ARMv8.2-A or later
187 @tab Enable the Reliability, Availability and Serviceability
189 @item @code{rcpc} @tab ARMv8.2-A @tab ARMv8.3-A or later
190 @tab Enable the weak release consistency extension.
191 @item @code{rdma} @tab ARMv8-A @tab ARMv8.1-A or later
192 @tab Enable ARMv8.1 Advanced SIMD extensions. This implies @code{simd}.
193 @item @code{simd} @tab ARMv8-A @tab ARMv8-A or later
194 @tab Enable Advanced SIMD extensions. This implies @code{fp}.
195 @item @code{sve} @tab ARMv8.2-A @tab No
196 @tab Enable the Scalable Vector Extensions. This implies @code{fp16},
197 @code{simd} and @code{compnum}.
198 @item @code{dotprod} @tab ARMv8.2-A @tab ARMv8.4-A or later
199 @tab Enable the Dot Product extension. This implies @code{simd}.
200 @item @code{fp16fml} @tab ARMv8.2-A @tab ARMv8.4-A or later
201 @tab Enable ARMv8.2 16-bit floating-point multiplication variant support.
202 This implies @code{fp16}.
203 @item @code{sb} @tab ARMv8-A @tab ARMv8.5-A or later
204 @tab Enable the speculation barrier instruction sb.
205 @item @code{predres} @tab ARMv8-A @tab ARMv8.5-A or later
206 @tab Enable the Execution and Data and Prediction instructions.
207 @item @code{rng} @tab ARMv8.5-A @tab No
208 @tab Enable ARMv8.5-A random number instructions.
209 @item @code{ssbs} @tab ARMv8-A @tab ARMv8.5-A or later
210 @tab Enable Speculative Store Bypassing Safe state read and write.
211 @item @code{memtag} @tab ARMv8.5-A @tab No
212 @tab Enable ARMv8.5-A Memory Tagging Extensions.
213 @item @code{tme} @tab ARMv8-A @tab No
214 @tab Enable Transactional Memory Extensions.
215 @item @code{sve2} @tab ARMv8-A @tab No
216 @tab Enable the SVE2 Extension.
217 @item @code{sve2-bitperm} @tab ARMv8-A @tab No
218 @tab Enable SVE2 BITPERM Extension.
219 @item @code{sve2-sm4} @tab ARMv8-A @tab No
220 @tab Enable SVE2 SM4 Extension.
221 @item @code{sve2-aes} @tab ARMv8-A @tab No
222 @tab Enable SVE2 AES Extension. This also enables the .Q->.B form of the
223 @code{pmullt} and @code{pmullb} instructions.
224 @item @code{sve2-sha3} @tab ARMv8-A @tab No
225 @tab Enable SVE2 SHA3 Extension.
231 * AArch64-Chars:: Special Characters
232 * AArch64-Regs:: Register Names
233 * AArch64-Relocations:: Relocations
237 @subsection Special Characters
239 @cindex line comment character, AArch64
240 @cindex AArch64 line comment character
241 The presence of a @samp{//} on a line indicates the start of a comment
242 that extends to the end of the current line. If a @samp{#} appears as
243 the first character of a line, the whole line is treated as a comment.
245 @cindex line separator, AArch64
246 @cindex statement separator, AArch64
247 @cindex AArch64 line separator
248 The @samp{;} character can be used instead of a newline to separate
251 @cindex immediate character, AArch64
252 @cindex AArch64 immediate character
253 The @samp{#} can be optionally used to indicate immediate operands.
256 @subsection Register Names
258 @cindex AArch64 register names
259 @cindex register names, AArch64
260 Please refer to the section @samp{4.4 Register Names} of
261 @samp{ARMv8 Instruction Set Overview}, which is available at
262 @uref{http://infocenter.arm.com}.
264 @node AArch64-Relocations
265 @subsection Relocations
267 @cindex relocations, AArch64
268 @cindex AArch64 relocations
269 @cindex MOVN, MOVZ and MOVK group relocations, AArch64
270 Relocations for @samp{MOVZ} and @samp{MOVK} instructions can be generated
271 by prefixing the label with @samp{#:abs_g2:} etc.
272 For example to load the 48-bit absolute address of @var{foo} into x0:
275 movz x0, #:abs_g2:foo // bits 32-47, overflow check
276 movk x0, #:abs_g1_nc:foo // bits 16-31, no overflow check
277 movk x0, #:abs_g0_nc:foo // bits 0-15, no overflow check
280 @cindex ADRP, ADD, LDR/STR group relocations, AArch64
281 Relocations for @samp{ADRP}, and @samp{ADD}, @samp{LDR} or @samp{STR}
282 instructions can be generated by prefixing the label with
283 @samp{:pg_hi21:} and @samp{#:lo12:} respectively.
285 For example to use 33-bit (+/-4GB) pc-relative addressing to
286 load the address of @var{foo} into x0:
289 adrp x0, :pg_hi21:foo
290 add x0, x0, #:lo12:foo
293 Or to load the value of @var{foo} into x0:
296 adrp x0, :pg_hi21:foo
297 ldr x0, [x0, #:lo12:foo]
300 Note that @samp{:pg_hi21:} is optional.
309 adrp x0, :pg_hi21:foo
312 @node AArch64 Floating Point
313 @section Floating Point
315 @cindex floating point, AArch64 (@sc{ieee})
316 @cindex AArch64 floating point (@sc{ieee})
317 The AArch64 architecture uses @sc{ieee} floating-point numbers.
319 @node AArch64 Directives
320 @section AArch64 Machine Directives
322 @cindex machine directives, AArch64
323 @cindex AArch64 machine directives
326 @c AAAAAAAAAAAAAAAAAAAAAAAAA
328 @cindex @code{.arch} directive, AArch64
329 @item .arch @var{name}
330 Select the target architecture. Valid values for @var{name} are the same as
331 for the @option{-march} command-line option.
333 Specifying @code{.arch} clears any previously selected architecture
336 @cindex @code{.arch_extension} directive, AArch64
337 @item .arch_extension @var{name}
338 Add or remove an architecture extension to the target architecture. Valid
339 values for @var{name} are the same as those accepted as architectural
340 extensions by the @option{-mcpu} command-line option.
342 @code{.arch_extension} may be used multiple times to add or remove extensions
343 incrementally to the architecture being compiled for.
345 @c BBBBBBBBBBBBBBBBBBBBBBBBBB
347 @cindex @code{.bss} directive, AArch64
349 This directive switches to the @code{.bss} section.
351 @c CCCCCCCCCCCCCCCCCCCCCCCCCC
353 @cindex @code{.cpu} directive, AArch64
354 @item .cpu @var{name}
355 Set the target processor. Valid values for @var{name} are the same as
356 those accepted by the @option{-mcpu=} command-line option.
358 @c DDDDDDDDDDDDDDDDDDDDDDDDDD
360 @cindex @code{.dword} directive, AArch64
361 @item .dword @var{expressions}
362 The @code{.dword} directive produces 64 bit values.
364 @c EEEEEEEEEEEEEEEEEEEEEEEEEE
366 @cindex @code{.even} directive, AArch64
368 The @code{.even} directive aligns the output on the next even byte
371 @c FFFFFFFFFFFFFFFFFFFFFFFFFF
373 @cindex @code{.float16} directive, AArch64
374 @item .float16 @var{value [,...,value_n]}
375 Place the half precision floating point representation of one or more
376 floating-point values into the current section.
377 The format used to encode the floating point values is always the
378 IEEE 754-2008 half precision floating point format.
380 @c GGGGGGGGGGGGGGGGGGGGGGGGGG
381 @c HHHHHHHHHHHHHHHHHHHHHHHHHH
382 @c IIIIIIIIIIIIIIIIIIIIIIIIII
384 @cindex @code{.inst} directive, AArch64
385 @item .inst @var{expressions}
386 Inserts the expressions into the output as if they were instructions,
389 @c JJJJJJJJJJJJJJJJJJJJJJJJJJ
390 @c KKKKKKKKKKKKKKKKKKKKKKKKKK
391 @c LLLLLLLLLLLLLLLLLLLLLLLLLL
393 @cindex @code{.ltorg} directive, AArch64
395 This directive causes the current contents of the literal pool to be
396 dumped into the current section (which is assumed to be the .text
397 section) at the current location (aligned to a word boundary).
398 GAS maintains a separate literal pool for each section and each
399 sub-section. The @code{.ltorg} directive will only affect the literal
400 pool of the current section and sub-section. At the end of assembly
401 all remaining, un-empty literal pools will automatically be dumped.
403 Note - older versions of GAS would dump the current literal
404 pool any time a section change occurred. This is no longer done, since
405 it prevents accurate control of the placement of literal pools.
407 @c MMMMMMMMMMMMMMMMMMMMMMMMMM
409 @c NNNNNNNNNNNNNNNNNNNNNNNNNN
410 @c OOOOOOOOOOOOOOOOOOOOOOOOOO
412 @c PPPPPPPPPPPPPPPPPPPPPPPPPP
414 @cindex @code{.pool} directive, AArch64
416 This is a synonym for .ltorg.
418 @c QQQQQQQQQQQQQQQQQQQQQQQQQQ
419 @c RRRRRRRRRRRRRRRRRRRRRRRRRR
421 @cindex @code{.req} directive, AArch64
422 @item @var{name} .req @var{register name}
423 This creates an alias for @var{register name} called @var{name}. For
430 ip0, ip1, lr and fp are automatically defined to
431 alias to X16, X17, X30 and X29 respectively.
433 @c SSSSSSSSSSSSSSSSSSSSSSSSSS
435 @c TTTTTTTTTTTTTTTTTTTTTTTTTT
437 @cindex @code{.tlsdescadd} directive, AArch64
438 @item @code{.tlsdescadd}
439 Emits a TLSDESC_ADD reloc on the next instruction.
441 @cindex @code{.tlsdesccall} directive, AArch64
442 @item @code{.tlsdesccall}
443 Emits a TLSDESC_CALL reloc on the next instruction.
445 @cindex @code{.tlsdescldr} directive, AArch64
446 @item @code{.tlsdescldr}
447 Emits a TLSDESC_LDR reloc on the next instruction.
449 @c UUUUUUUUUUUUUUUUUUUUUUUUUU
451 @cindex @code{.unreq} directive, AArch64
452 @item .unreq @var{alias-name}
453 This undefines a register alias which was previously defined using the
454 @code{req} directive. For example:
461 An error occurs if the name is undefined. Note - this pseudo op can
462 be used to delete builtin in register name aliases (eg 'w0'). This
463 should only be done if it is really necessary.
465 @c VVVVVVVVVVVVVVVVVVVVVVVVVV
467 @cindex @code{.variant_pcs} directive, AArch64
468 @item .variant_pcs @var{symbol}
469 This directive marks @var{symbol} referencing a function that may
470 follow a variant procedure call standard with different register
471 usage convention from the base procedure call standard.
473 @c WWWWWWWWWWWWWWWWWWWWWWWWWW
474 @c XXXXXXXXXXXXXXXXXXXXXXXXXX
476 @cindex @code{.xword} directive, AArch64
477 @item .xword @var{expressions}
478 The @code{.xword} directive produces 64 bit values. This is the same
479 as the @code{.dword} directive.
481 @c YYYYYYYYYYYYYYYYYYYYYYYYYY
482 @c ZZZZZZZZZZZZZZZZZZZZZZZZZZ
484 @cindex @code{.cfi_b_key_frame} directive, AArch64
485 @item @code{.cfi_b_key_frame}
486 The @code{.cfi_b_key_frame} directive inserts a 'B' character into the CIE
487 corresponding to the current frame's FDE, meaning that its return address has
488 been signed with the B-key. If two frames are signed with differing keys then
489 they will not share the same CIE. This information is intended to be used by
490 the stack unwinder in order to properly authenticate return addresses.
494 @node AArch64 Opcodes
497 @cindex AArch64 opcodes
498 @cindex opcodes for AArch64
499 GAS implements all the standard AArch64 opcodes. It also
500 implements several pseudo opcodes, including several synthetic load
505 @cindex @code{LDR reg,=<expr>} pseudo op, AArch64
508 ldr <register> , =<expression>
511 The constant expression will be placed into the nearest literal pool (if it not
512 already there) and a PC-relative LDR instruction will be generated.
516 For more information on the AArch64 instruction set and assembly language
517 notation, see @samp{ARMv8 Instruction Set Overview} available at
518 @uref{http://infocenter.arm.com}.
521 @node AArch64 Mapping Symbols
522 @section Mapping Symbols
524 The AArch64 ELF specification requires that special symbols be inserted
525 into object files to mark certain features:
531 At the start of a region of code containing AArch64 instructions.
535 At the start of a region of data.