Commit | Line | Data |
---|---|---|
82704155 | 1 | @c Copyright (C) 1996-2019 Free Software Foundation, Inc. |
252b5132 RH |
2 | @c This is part of the GAS manual. |
3 | @c For copying conditions, see the file as.texinfo. | |
4 | ||
5 | @ifset GENERIC | |
6 | @page | |
7 | @node ARM-Dependent | |
8 | @chapter ARM Dependent Features | |
9 | @end ifset | |
10 | ||
11 | @ifclear GENERIC | |
12 | @node Machine Dependencies | |
13 | @chapter ARM Dependent Features | |
14 | @end ifclear | |
15 | ||
16 | @cindex ARM support | |
17 | @cindex Thumb support | |
18 | @menu | |
19 | * ARM Options:: Options | |
20 | * ARM Syntax:: Syntax | |
21 | * ARM Floating Point:: Floating Point | |
22 | * ARM Directives:: ARM Machine Directives | |
23 | * ARM Opcodes:: Opcodes | |
6057a28f | 24 | * ARM Mapping Symbols:: Mapping Symbols |
7da4f750 | 25 | * ARM Unwinding Tutorial:: Unwinding |
252b5132 RH |
26 | @end menu |
27 | ||
28 | @node ARM Options | |
29 | @section Options | |
30 | @cindex ARM options (none) | |
31 | @cindex options for ARM (none) | |
adcf07e6 | 32 | |
252b5132 | 33 | @table @code |
adcf07e6 | 34 | |
a05a5b64 | 35 | @cindex @code{-mcpu=} command-line option, ARM |
92081f48 | 36 | @item -mcpu=@var{processor}[+@var{extension}@dots{}] |
252b5132 RH |
37 | This option specifies the target processor. The assembler will issue an |
38 | error message if an attempt is made to assemble an instruction which | |
03b1477f | 39 | will not execute on the target processor. The following processor names are |
34bca508 | 40 | recognized: |
03b1477f RE |
41 | @code{arm1}, |
42 | @code{arm2}, | |
43 | @code{arm250}, | |
44 | @code{arm3}, | |
45 | @code{arm6}, | |
46 | @code{arm60}, | |
47 | @code{arm600}, | |
48 | @code{arm610}, | |
49 | @code{arm620}, | |
50 | @code{arm7}, | |
51 | @code{arm7m}, | |
52 | @code{arm7d}, | |
53 | @code{arm7dm}, | |
54 | @code{arm7di}, | |
55 | @code{arm7dmi}, | |
56 | @code{arm70}, | |
57 | @code{arm700}, | |
58 | @code{arm700i}, | |
59 | @code{arm710}, | |
60 | @code{arm710t}, | |
61 | @code{arm720}, | |
62 | @code{arm720t}, | |
63 | @code{arm740t}, | |
64 | @code{arm710c}, | |
65 | @code{arm7100}, | |
66 | @code{arm7500}, | |
67 | @code{arm7500fe}, | |
68 | @code{arm7t}, | |
69 | @code{arm7tdmi}, | |
1ff4677c | 70 | @code{arm7tdmi-s}, |
03b1477f RE |
71 | @code{arm8}, |
72 | @code{arm810}, | |
73 | @code{strongarm}, | |
74 | @code{strongarm1}, | |
75 | @code{strongarm110}, | |
76 | @code{strongarm1100}, | |
77 | @code{strongarm1110}, | |
78 | @code{arm9}, | |
79 | @code{arm920}, | |
80 | @code{arm920t}, | |
81 | @code{arm922t}, | |
82 | @code{arm940t}, | |
83 | @code{arm9tdmi}, | |
7fac0536 NC |
84 | @code{fa526} (Faraday FA526 processor), |
85 | @code{fa626} (Faraday FA626 processor), | |
03b1477f | 86 | @code{arm9e}, |
7de9afa2 | 87 | @code{arm926e}, |
1ff4677c | 88 | @code{arm926ej-s}, |
03b1477f RE |
89 | @code{arm946e-r0}, |
90 | @code{arm946e}, | |
db8ac8f9 | 91 | @code{arm946e-s}, |
03b1477f RE |
92 | @code{arm966e-r0}, |
93 | @code{arm966e}, | |
db8ac8f9 PB |
94 | @code{arm966e-s}, |
95 | @code{arm968e-s}, | |
03b1477f | 96 | @code{arm10t}, |
db8ac8f9 | 97 | @code{arm10tdmi}, |
03b1477f RE |
98 | @code{arm10e}, |
99 | @code{arm1020}, | |
100 | @code{arm1020t}, | |
7de9afa2 | 101 | @code{arm1020e}, |
db8ac8f9 | 102 | @code{arm1022e}, |
1ff4677c | 103 | @code{arm1026ej-s}, |
4a58c4bd NC |
104 | @code{fa606te} (Faraday FA606TE processor), |
105 | @code{fa616te} (Faraday FA616TE processor), | |
7fac0536 | 106 | @code{fa626te} (Faraday FA626TE processor), |
4a58c4bd | 107 | @code{fmp626} (Faraday FMP626 processor), |
7fac0536 | 108 | @code{fa726te} (Faraday FA726TE processor), |
1ff4677c RE |
109 | @code{arm1136j-s}, |
110 | @code{arm1136jf-s}, | |
db8ac8f9 PB |
111 | @code{arm1156t2-s}, |
112 | @code{arm1156t2f-s}, | |
0dd132b6 NC |
113 | @code{arm1176jz-s}, |
114 | @code{arm1176jzf-s}, | |
115 | @code{mpcore}, | |
116 | @code{mpcorenovfp}, | |
b38f9f31 | 117 | @code{cortex-a5}, |
c90460e4 | 118 | @code{cortex-a7}, |
62b3e311 | 119 | @code{cortex-a8}, |
15290f0a | 120 | @code{cortex-a9}, |
dbb1f804 | 121 | @code{cortex-a15}, |
ed5491b9 | 122 | @code{cortex-a17}, |
6735952f | 123 | @code{cortex-a32}, |
43cdc0a8 | 124 | @code{cortex-a35}, |
4469186b | 125 | @code{cortex-a53}, |
15a7695f | 126 | @code{cortex-a55}, |
4469186b KT |
127 | @code{cortex-a57}, |
128 | @code{cortex-a72}, | |
362a3eba | 129 | @code{cortex-a73}, |
15a7695f | 130 | @code{cortex-a75}, |
7ebd1359 | 131 | @code{cortex-a76}, |
0535e5d7 DZ |
132 | @code{cortex-a76ae}, |
133 | @code{cortex-a77}, | |
ef8df4ca | 134 | @code{ares}, |
62b3e311 | 135 | @code{cortex-r4}, |
307c948d | 136 | @code{cortex-r4f}, |
70a8bc5b | 137 | @code{cortex-r5}, |
138 | @code{cortex-r7}, | |
5f474010 | 139 | @code{cortex-r8}, |
0cda1e19 | 140 | @code{cortex-r52}, |
0535e5d7 | 141 | @code{cortex-m35p}, |
b19ea8d2 | 142 | @code{cortex-m33}, |
ce1b0a45 | 143 | @code{cortex-m23}, |
a715796b | 144 | @code{cortex-m7}, |
7ef07ba0 | 145 | @code{cortex-m4}, |
62b3e311 | 146 | @code{cortex-m3}, |
5b19eaba NC |
147 | @code{cortex-m1}, |
148 | @code{cortex-m0}, | |
ce32bd10 | 149 | @code{cortex-m0plus}, |
246496bb | 150 | @code{exynos-m1}, |
ea0d6bb9 PT |
151 | @code{marvell-pj4}, |
152 | @code{marvell-whitney}, | |
83f43c83 | 153 | @code{neoverse-n1}, |
ea0d6bb9 PT |
154 | @code{xgene1}, |
155 | @code{xgene2}, | |
03b1477f RE |
156 | @code{ep9312} (ARM920 with Cirrus Maverick coprocessor), |
157 | @code{i80200} (Intel XScale processor) | |
334fe02b | 158 | @code{iwmmxt} (Intel XScale processor with Wireless MMX technology coprocessor) |
03b1477f | 159 | and |
34bca508 | 160 | @code{xscale}. |
03b1477f RE |
161 | The special name @code{all} may be used to allow the |
162 | assembler to accept instructions valid for any ARM processor. | |
163 | ||
34bca508 L |
164 | In addition to the basic instruction set, the assembler can be told to |
165 | accept various extension mnemonics that extend the processor using the | |
03b1477f | 166 | co-processor instruction space. For example, @code{-mcpu=arm920+maverick} |
34bca508 | 167 | is equivalent to specifying @code{-mcpu=ep9312}. |
69133863 | 168 | |
34bca508 | 169 | Multiple extensions may be specified, separated by a @code{+}. The |
69133863 MGD |
170 | extensions should be specified in ascending alphabetical order. |
171 | ||
34bca508 | 172 | Some extensions may be restricted to particular architectures; this is |
60e5ef9f MGD |
173 | documented in the list of extensions below. |
174 | ||
34bca508 L |
175 | Extension mnemonics may also be removed from those the assembler accepts. |
176 | This is done be prepending @code{no} to the option that adds the extension. | |
177 | Extensions that are removed should be listed after all extensions which have | |
178 | been added, again in ascending alphabetical order. For example, | |
69133863 MGD |
179 | @code{-mcpu=ep9312+nomaverick} is equivalent to specifying @code{-mcpu=arm920}. |
180 | ||
181 | ||
eea54501 | 182 | The following extensions are currently supported: |
aab2c27d | 183 | @code{bf16} (BFloat16 extensions for v8.6-A architecture), |
616ce08e | 184 | @code{i8mm} (Int8 Matrix Multiply extensions for v8.6-A architecture), |
ea0d6bb9 | 185 | @code{crc} |
bca38921 | 186 | @code{crypto} (Cryptography Extensions for v8-A architecture, implies @code{fp+simd}), |
c604a79a | 187 | @code{dotprod} (Dot Product Extensions for v8.2-A architecture, implies @code{fp+simd}), |
bca38921 | 188 | @code{fp} (Floating Point Extensions for v8-A architecture), |
01f48020 TC |
189 | @code{fp16} (FP16 Extensions for v8.2-A architecture, implies @code{fp}), |
190 | @code{fp16fml} (FP16 Floating Point Multiplication Variant Extensions for v8.2-A architecture, implies @code{fp16}), | |
bca38921 | 191 | @code{idiv} (Integer Divide Extensions for v7-A and v7-R architectures), |
69133863 MGD |
192 | @code{iwmmxt}, |
193 | @code{iwmmxt2}, | |
ea0d6bb9 | 194 | @code{xscale}, |
69133863 | 195 | @code{maverick}, |
ea0d6bb9 PT |
196 | @code{mp} (Multiprocessing Extensions for v7-A and v7-R |
197 | architectures), | |
b2a5fbdc | 198 | @code{os} (Operating System for v6M architecture), |
dad0c3bf SD |
199 | @code{predres} (Execution and Data Prediction Restriction Instruction for |
200 | v8-A architectures, added by default from v8.5-A), | |
7fadb25d SD |
201 | @code{sb} (Speculation Barrier Instruction for v8-A architectures, added by |
202 | default from v8.5-A), | |
f4c65163 | 203 | @code{sec} (Security Extensions for v6K and v7-A architectures), |
bca38921 | 204 | @code{simd} (Advanced SIMD Extensions for v8-A architecture, implies @code{fp}), |
34bca508 | 205 | @code{virt} (Virtualization Extensions for v7-A architecture, implies |
90ec0d68 | 206 | @code{idiv}), |
33eaf5de | 207 | @code{pan} (Privileged Access Never Extensions for v8-A architecture), |
4d1464f2 MW |
208 | @code{ras} (Reliability, Availability and Serviceability extensions |
209 | for v8-A architecture), | |
d6b4b13e MW |
210 | @code{rdma} (ARMv8.1 Advanced SIMD extensions for v8-A architecture, implies |
211 | @code{simd}) | |
03b1477f | 212 | and |
69133863 | 213 | @code{xscale}. |
03b1477f | 214 | |
a05a5b64 | 215 | @cindex @code{-march=} command-line option, ARM |
92081f48 | 216 | @item -march=@var{architecture}[+@var{extension}@dots{}] |
252b5132 RH |
217 | This option specifies the target architecture. The assembler will issue |
218 | an error message if an attempt is made to assemble an instruction which | |
34bca508 L |
219 | will not execute on the target architecture. The following architecture |
220 | names are recognized: | |
03b1477f RE |
221 | @code{armv1}, |
222 | @code{armv2}, | |
223 | @code{armv2a}, | |
224 | @code{armv2s}, | |
225 | @code{armv3}, | |
226 | @code{armv3m}, | |
227 | @code{armv4}, | |
228 | @code{armv4xm}, | |
229 | @code{armv4t}, | |
230 | @code{armv4txm}, | |
231 | @code{armv5}, | |
232 | @code{armv5t}, | |
233 | @code{armv5txm}, | |
234 | @code{armv5te}, | |
09d92015 | 235 | @code{armv5texp}, |
c5f98204 | 236 | @code{armv6}, |
1ddd7f43 | 237 | @code{armv6j}, |
0dd132b6 NC |
238 | @code{armv6k}, |
239 | @code{armv6z}, | |
f33026a9 | 240 | @code{armv6kz}, |
b2a5fbdc MGD |
241 | @code{armv6-m}, |
242 | @code{armv6s-m}, | |
62b3e311 | 243 | @code{armv7}, |
c450d570 | 244 | @code{armv7-a}, |
c9fb6e58 | 245 | @code{armv7ve}, |
c450d570 PB |
246 | @code{armv7-r}, |
247 | @code{armv7-m}, | |
9e3c6df6 | 248 | @code{armv7e-m}, |
bca38921 | 249 | @code{armv8-a}, |
a5932920 | 250 | @code{armv8.1-a}, |
56a1b672 | 251 | @code{armv8.2-a}, |
a12fd8e1 | 252 | @code{armv8.3-a}, |
ced40572 | 253 | @code{armv8-r}, |
dec41383 | 254 | @code{armv8.4-a}, |
23f233a5 | 255 | @code{armv8.5-a}, |
34ef62f4 AV |
256 | @code{armv8-m.base}, |
257 | @code{armv8-m.main}, | |
e0991585 | 258 | @code{armv8.1-m.main}, |
aab2c27d | 259 | @code{armv8.6-a}, |
34ef62f4 | 260 | @code{iwmmxt}, |
ea0d6bb9 | 261 | @code{iwmmxt2} |
03b1477f RE |
262 | and |
263 | @code{xscale}. | |
264 | If both @code{-mcpu} and | |
265 | @code{-march} are specified, the assembler will use | |
266 | the setting for @code{-mcpu}. | |
267 | ||
34ef62f4 AV |
268 | The architecture option can be extended with a set extension options. These |
269 | extensions are context sensitive, i.e. the same extension may mean different | |
270 | things when used with different architectures. When used together with a | |
271 | @code{-mfpu} option, the union of both feature enablement is taken. | |
272 | See their availability and meaning below: | |
273 | ||
274 | For @code{armv5te}, @code{armv5texp}, @code{armv5tej}, @code{armv6}, @code{armv6j}, @code{armv6k}, @code{armv6z}, @code{armv6kz}, @code{armv6zk}, @code{armv6t2}, @code{armv6kt2} and @code{armv6zt2}: | |
275 | ||
276 | @code{+fp}: Enables VFPv2 instructions. | |
277 | @code{+nofp}: Disables all FPU instrunctions. | |
278 | ||
279 | For @code{armv7}: | |
280 | ||
281 | @code{+fp}: Enables VFPv3 instructions with 16 double-word registers. | |
282 | @code{+nofp}: Disables all FPU instructions. | |
283 | ||
284 | For @code{armv7-a}: | |
285 | ||
286 | @code{+fp}: Enables VFPv3 instructions with 16 double-word registers. | |
287 | @code{+vfpv3-d16}: Alias for @code{+fp}. | |
288 | @code{+vfpv3}: Enables VFPv3 instructions with 32 double-word registers. | |
289 | @code{+vfpv3-d16-fp16}: Enables VFPv3 with half precision floating-point | |
290 | conversion instructions and 16 double-word registers. | |
291 | @code{+vfpv3-fp16}: Enables VFPv3 with half precision floating-point conversion | |
292 | instructions and 32 double-word registers. | |
293 | @code{+vfpv4-d16}: Enables VFPv4 instructions with 16 double-word registers. | |
294 | @code{+vfpv4}: Enables VFPv4 instructions with 32 double-word registers. | |
295 | @code{+simd}: Enables VFPv3 and NEONv1 instructions with 32 double-word | |
296 | registers. | |
297 | @code{+neon}: Alias for @code{+simd}. | |
298 | @code{+neon-vfpv3}: Alias for @code{+simd}. | |
299 | @code{+neon-fp16}: Enables VFPv3, half precision floating-point conversion and | |
300 | NEONv1 instructions with 32 double-word registers. | |
301 | @code{+neon-vfpv4}: Enables VFPv4 and NEONv1 with Fused-MAC instructions and 32 | |
302 | double-word registers. | |
303 | @code{+mp}: Enables Multiprocessing Extensions. | |
304 | @code{+sec}: Enables Security Extensions. | |
305 | @code{+nofp}: Disables all FPU and NEON instructions. | |
306 | @code{+nosimd}: Disables all NEON instructions. | |
307 | ||
308 | For @code{armv7ve}: | |
309 | ||
310 | @code{+fp}: Enables VFPv4 instructions with 16 double-word registers. | |
311 | @code{+vfpv4-d16}: Alias for @code{+fp}. | |
312 | @code{+vfpv3-d16}: Enables VFPv3 instructions with 16 double-word registers. | |
313 | @code{+vfpv3}: Enables VFPv3 instructions with 32 double-word registers. | |
314 | @code{+vfpv3-d16-fp16}: Enables VFPv3 with half precision floating-point | |
315 | conversion instructions and 16 double-word registers. | |
316 | @code{+vfpv3-fp16}: Enables VFPv3 with half precision floating-point conversion | |
317 | instructions and 32 double-word registers. | |
318 | @code{+vfpv4}: Enables VFPv4 instructions with 32 double-word registers. | |
319 | @code{+simd}: Enables VFPv4 and NEONv1 with Fused-MAC instructions and 32 | |
320 | double-word registers. | |
321 | @code{+neon-vfpv4}: Alias for @code{+simd}. | |
322 | @code{+neon}: Enables VFPv3 and NEONv1 instructions with 32 double-word | |
323 | registers. | |
324 | @code{+neon-vfpv3}: Alias for @code{+neon}. | |
325 | @code{+neon-fp16}: Enables VFPv3, half precision floating-point conversion and | |
326 | NEONv1 instructions with 32 double-word registers. | |
327 | double-word registers. | |
328 | @code{+nofp}: Disables all FPU and NEON instructions. | |
329 | @code{+nosimd}: Disables all NEON instructions. | |
330 | ||
331 | For @code{armv7-r}: | |
332 | ||
333 | @code{+fp.sp}: Enables single-precision only VFPv3 instructions with 16 | |
334 | double-word registers. | |
335 | @code{+vfpv3xd}: Alias for @code{+fp.sp}. | |
336 | @code{+fp}: Enables VFPv3 instructions with 16 double-word registers. | |
337 | @code{+vfpv3-d16}: Alias for @code{+fp}. | |
338 | @code{+vfpv3xd-fp16}: Enables single-precision only VFPv3 and half | |
339 | floating-point conversion instructions with 16 double-word registers. | |
340 | @code{+vfpv3-d16-fp16}: Enables VFPv3 and half precision floating-point | |
341 | conversion instructions with 16 double-word registers. | |
342 | @code{+idiv}: Enables integer division instructions in ARM mode. | |
343 | @code{+nofp}: Disables all FPU instructions. | |
344 | ||
345 | For @code{armv7e-m}: | |
346 | ||
347 | @code{+fp}: Enables single-precision only VFPv4 instructions with 16 | |
348 | double-word registers. | |
349 | @code{+vfpvf4-sp-d16}: Alias for @code{+fp}. | |
350 | @code{+fpv5}: Enables single-precision only VFPv5 instructions with 16 | |
351 | double-word registers. | |
352 | @code{+fp.dp}: Enables VFPv5 instructions with 16 double-word registers. | |
353 | @code{+fpv5-d16"}: Alias for @code{+fp.dp}. | |
354 | @code{+nofp}: Disables all FPU instructions. | |
355 | ||
356 | For @code{armv8-m.main}: | |
357 | ||
358 | @code{+dsp}: Enables DSP Extension. | |
359 | @code{+fp}: Enables single-precision only VFPv5 instructions with 16 | |
360 | double-word registers. | |
361 | @code{+fp.dp}: Enables VFPv5 instructions with 16 double-word registers. | |
362 | @code{+nofp}: Disables all FPU instructions. | |
363 | @code{+nodsp}: Disables DSP Extension. | |
364 | ||
e0991585 AV |
365 | For @code{armv8.1-m.main}: |
366 | ||
367 | @code{+dsp}: Enables DSP Extension. | |
368 | @code{+fp}: Enables single and half precision scalar Floating Point Extensions | |
369 | for Armv8.1-M Mainline with 16 double-word registers. | |
370 | @code{+fp.dp}: Enables double precision scalar Floating Point Extensions for | |
371 | Armv8.1-M Mainline, implies @code{+fp}. | |
a7ad558c AV |
372 | @code{+mve}: Enables integer only M-profile Vector Extension for |
373 | Armv8.1-M Mainline, implies @code{+dsp}. | |
374 | @code{+mve.fp}: Enables Floating Point M-profile Vector Extension for | |
375 | Armv8.1-M Mainline, implies @code{+mve} and @code{+fp}. | |
e0991585 AV |
376 | @code{+nofp}: Disables all FPU instructions. |
377 | @code{+nodsp}: Disables DSP Extension. | |
a7ad558c | 378 | @code{+nomve}: Disables all M-profile Vector Extensions. |
e0991585 | 379 | |
34ef62f4 AV |
380 | For @code{armv8-a}: |
381 | ||
382 | @code{+crc}: Enables CRC32 Extension. | |
383 | @code{+simd}: Enables VFP and NEON for Armv8-A. | |
384 | @code{+crypto}: Enables Cryptography Extensions for Armv8-A, implies | |
385 | @code{+simd}. | |
386 | @code{+sb}: Enables Speculation Barrier Instruction for Armv8-A. | |
387 | @code{+predres}: Enables Execution and Data Prediction Restriction Instruction | |
388 | for Armv8-A. | |
389 | @code{+nofp}: Disables all FPU, NEON and Cryptography Extensions. | |
390 | @code{+nocrypto}: Disables Cryptography Extensions. | |
391 | ||
392 | For @code{armv8.1-a}: | |
393 | ||
394 | @code{+simd}: Enables VFP and NEON for Armv8.1-A. | |
395 | @code{+crypto}: Enables Cryptography Extensions for Armv8-A, implies | |
396 | @code{+simd}. | |
397 | @code{+sb}: Enables Speculation Barrier Instruction for Armv8-A. | |
398 | @code{+predres}: Enables Execution and Data Prediction Restriction Instruction | |
399 | for Armv8-A. | |
400 | @code{+nofp}: Disables all FPU, NEON and Cryptography Extensions. | |
401 | @code{+nocrypto}: Disables Cryptography Extensions. | |
402 | ||
403 | For @code{armv8.2-a} and @code{armv8.3-a}: | |
404 | ||
405 | @code{+simd}: Enables VFP and NEON for Armv8.1-A. | |
406 | @code{+fp16}: Enables FP16 Extension for Armv8.2-A, implies @code{+simd}. | |
407 | @code{+fp16fml}: Enables FP16 Floating Point Multiplication Variant Extensions | |
408 | for Armv8.2-A, implies @code{+fp16}. | |
409 | @code{+crypto}: Enables Cryptography Extensions for Armv8-A, implies | |
410 | @code{+simd}. | |
411 | @code{+dotprod}: Enables Dot Product Extensions for Armv8.2-A, implies | |
412 | @code{+simd}. | |
413 | @code{+sb}: Enables Speculation Barrier Instruction for Armv8-A. | |
414 | @code{+predres}: Enables Execution and Data Prediction Restriction Instruction | |
415 | for Armv8-A. | |
416 | @code{+nofp}: Disables all FPU, NEON, Cryptography and Dot Product Extensions. | |
417 | @code{+nocrypto}: Disables Cryptography Extensions. | |
418 | ||
419 | For @code{armv8.4-a}: | |
420 | ||
421 | @code{+simd}: Enables VFP and NEON for Armv8.1-A and Dot Product Extensions for | |
422 | Armv8.2-A. | |
423 | @code{+fp16}: Enables FP16 Floating Point and Floating Point Multiplication | |
424 | Variant Extensions for Armv8.2-A, implies @code{+simd}. | |
425 | @code{+crypto}: Enables Cryptography Extensions for Armv8-A, implies | |
426 | @code{+simd}. | |
427 | @code{+sb}: Enables Speculation Barrier Instruction for Armv8-A. | |
428 | @code{+predres}: Enables Execution and Data Prediction Restriction Instruction | |
429 | for Armv8-A. | |
430 | @code{+nofp}: Disables all FPU, NEON, Cryptography and Dot Product Extensions. | |
431 | @code{+nocryptp}: Disables Cryptography Extensions. | |
432 | ||
433 | For @code{armv8.5-a}: | |
434 | ||
435 | @code{+simd}: Enables VFP and NEON for Armv8.1-A and Dot Product Extensions for | |
436 | Armv8.2-A. | |
437 | @code{+fp16}: Enables FP16 Floating Point and Floating Point Multiplication | |
438 | Variant Extensions for Armv8.2-A, implies @code{+simd}. | |
439 | @code{+crypto}: Enables Cryptography Extensions for Armv8-A, implies | |
440 | @code{+simd}. | |
441 | @code{+nofp}: Disables all FPU, NEON, Cryptography and Dot Product Extensions. | |
442 | @code{+nocryptp}: Disables Cryptography Extensions. | |
443 | ||
03b1477f | 444 | |
a05a5b64 | 445 | @cindex @code{-mfpu=} command-line option, ARM |
03b1477f RE |
446 | @item -mfpu=@var{floating-point-format} |
447 | ||
448 | This option specifies the floating point format to assemble for. The | |
449 | assembler will issue an error message if an attempt is made to assemble | |
34bca508 | 450 | an instruction which will not execute on the target floating point unit. |
03b1477f RE |
451 | The following format options are recognized: |
452 | @code{softfpa}, | |
453 | @code{fpe}, | |
bc89618b RE |
454 | @code{fpe2}, |
455 | @code{fpe3}, | |
03b1477f RE |
456 | @code{fpa}, |
457 | @code{fpa10}, | |
458 | @code{fpa11}, | |
459 | @code{arm7500fe}, | |
460 | @code{softvfp}, | |
461 | @code{softvfp+vfp}, | |
462 | @code{vfp}, | |
463 | @code{vfp10}, | |
464 | @code{vfp10-r0}, | |
465 | @code{vfp9}, | |
466 | @code{vfpxd}, | |
62f3b8c8 PB |
467 | @code{vfpv2}, |
468 | @code{vfpv3}, | |
469 | @code{vfpv3-fp16}, | |
470 | @code{vfpv3-d16}, | |
471 | @code{vfpv3-d16-fp16}, | |
472 | @code{vfpv3xd}, | |
473 | @code{vfpv3xd-d16}, | |
474 | @code{vfpv4}, | |
475 | @code{vfpv4-d16}, | |
f0cd0667 | 476 | @code{fpv4-sp-d16}, |
a715796b TG |
477 | @code{fpv5-sp-d16}, |
478 | @code{fpv5-d16}, | |
bca38921 | 479 | @code{fp-armv8}, |
09d92015 MM |
480 | @code{arm1020t}, |
481 | @code{arm1020e}, | |
b1cc4aeb | 482 | @code{arm1136jf-s}, |
62f3b8c8 PB |
483 | @code{maverick}, |
484 | @code{neon}, | |
d5e0ba9c RE |
485 | @code{neon-vfpv3}, |
486 | @code{neon-fp16}, | |
bca38921 MGD |
487 | @code{neon-vfpv4}, |
488 | @code{neon-fp-armv8}, | |
081e4c7d MW |
489 | @code{crypto-neon-fp-armv8}, |
490 | @code{neon-fp-armv8.1} | |
d6b4b13e | 491 | and |
081e4c7d | 492 | @code{crypto-neon-fp-armv8.1}. |
03b1477f RE |
493 | |
494 | In addition to determining which instructions are assembled, this option | |
495 | also affects the way in which the @code{.double} assembler directive behaves | |
496 | when assembling little-endian code. | |
497 | ||
34bca508 | 498 | The default is dependent on the processor selected. For Architecture 5 or |
d5e0ba9c | 499 | later, the default is to assemble for VFP instructions; for earlier |
03b1477f | 500 | architectures the default is to assemble for FPA instructions. |
adcf07e6 | 501 | |
5312fe52 BW |
502 | @cindex @code{-mfp16-format=} command-line option |
503 | @item -mfp16-format=@var{format} | |
504 | This option specifies the half-precision floating point format to use | |
505 | when assembling floating point numbers emitted by the @code{.float16} | |
506 | directive. | |
507 | The following format options are recognized: | |
508 | @code{ieee}, | |
509 | @code{alternative}. | |
510 | If @code{ieee} is specified then the IEEE 754-2008 half-precision floating | |
511 | point format is used, if @code{alternative} is specified then the Arm | |
512 | alternative half-precision format is used. If this option is set on the | |
513 | command line then the format is fixed and cannot be changed with | |
514 | the @code{float16_format} directive. If this value is not set then | |
515 | the IEEE 754-2008 format is used until the format is explicitly set with | |
516 | the @code{float16_format} directive. | |
517 | ||
a05a5b64 | 518 | @cindex @code{-mthumb} command-line option, ARM |
252b5132 | 519 | @item -mthumb |
03b1477f | 520 | This option specifies that the assembler should start assembling Thumb |
34bca508 | 521 | instructions; that is, it should behave as though the file starts with a |
03b1477f | 522 | @code{.code 16} directive. |
adcf07e6 | 523 | |
a05a5b64 | 524 | @cindex @code{-mthumb-interwork} command-line option, ARM |
252b5132 RH |
525 | @item -mthumb-interwork |
526 | This option specifies that the output generated by the assembler should | |
fc6141f0 NC |
527 | be marked as supporting interworking. It also affects the behaviour |
528 | of the @code{ADR} and @code{ADRL} pseudo opcodes. | |
adcf07e6 | 529 | |
a05a5b64 | 530 | @cindex @code{-mimplicit-it} command-line option, ARM |
52970753 NC |
531 | @item -mimplicit-it=never |
532 | @itemx -mimplicit-it=always | |
533 | @itemx -mimplicit-it=arm | |
534 | @itemx -mimplicit-it=thumb | |
535 | The @code{-mimplicit-it} option controls the behavior of the assembler when | |
536 | conditional instructions are not enclosed in IT blocks. | |
537 | There are four possible behaviors. | |
538 | If @code{never} is specified, such constructs cause a warning in ARM | |
539 | code and an error in Thumb-2 code. | |
540 | If @code{always} is specified, such constructs are accepted in both | |
541 | ARM and Thumb-2 code, where the IT instruction is added implicitly. | |
542 | If @code{arm} is specified, such constructs are accepted in ARM code | |
543 | and cause an error in Thumb-2 code. | |
544 | If @code{thumb} is specified, such constructs cause a warning in ARM | |
545 | code and are accepted in Thumb-2 code. If you omit this option, the | |
546 | behavior is equivalent to @code{-mimplicit-it=arm}. | |
e07e6e58 | 547 | |
a05a5b64 TP |
548 | @cindex @code{-mapcs-26} command-line option, ARM |
549 | @cindex @code{-mapcs-32} command-line option, ARM | |
5a5829dd NS |
550 | @item -mapcs-26 |
551 | @itemx -mapcs-32 | |
552 | These options specify that the output generated by the assembler should | |
252b5132 RH |
553 | be marked as supporting the indicated version of the Arm Procedure. |
554 | Calling Standard. | |
adcf07e6 | 555 | |
a05a5b64 | 556 | @cindex @code{-matpcs} command-line option, ARM |
077b8428 | 557 | @item -matpcs |
34bca508 | 558 | This option specifies that the output generated by the assembler should |
077b8428 NC |
559 | be marked as supporting the Arm/Thumb Procedure Calling Standard. If |
560 | enabled this option will cause the assembler to create an empty | |
561 | debugging section in the object file called .arm.atpcs. Debuggers can | |
562 | use this to determine the ABI being used by. | |
563 | ||
a05a5b64 | 564 | @cindex @code{-mapcs-float} command-line option, ARM |
252b5132 | 565 | @item -mapcs-float |
1be59579 | 566 | This indicates the floating point variant of the APCS should be |
252b5132 | 567 | used. In this variant floating point arguments are passed in FP |
550262c4 | 568 | registers rather than integer registers. |
adcf07e6 | 569 | |
a05a5b64 | 570 | @cindex @code{-mapcs-reentrant} command-line option, ARM |
252b5132 RH |
571 | @item -mapcs-reentrant |
572 | This indicates that the reentrant variant of the APCS should be used. | |
573 | This variant supports position independent code. | |
adcf07e6 | 574 | |
a05a5b64 | 575 | @cindex @code{-mfloat-abi=} command-line option, ARM |
33a392fb PB |
576 | @item -mfloat-abi=@var{abi} |
577 | This option specifies that the output generated by the assembler should be | |
578 | marked as using specified floating point ABI. | |
579 | The following values are recognized: | |
580 | @code{soft}, | |
581 | @code{softfp} | |
582 | and | |
583 | @code{hard}. | |
584 | ||
a05a5b64 | 585 | @cindex @code{-eabi=} command-line option, ARM |
d507cf36 PB |
586 | @item -meabi=@var{ver} |
587 | This option specifies which EABI version the produced object files should | |
588 | conform to. | |
b45619c0 | 589 | The following values are recognized: |
3a4a14e9 PB |
590 | @code{gnu}, |
591 | @code{4} | |
d507cf36 | 592 | and |
3a4a14e9 | 593 | @code{5}. |
d507cf36 | 594 | |
a05a5b64 | 595 | @cindex @code{-EB} command-line option, ARM |
252b5132 RH |
596 | @item -EB |
597 | This option specifies that the output generated by the assembler should | |
598 | be marked as being encoded for a big-endian processor. | |
adcf07e6 | 599 | |
080bb7bb NC |
600 | Note: If a program is being built for a system with big-endian data |
601 | and little-endian instructions then it should be assembled with the | |
602 | @option{-EB} option, (all of it, code and data) and then linked with | |
603 | the @option{--be8} option. This will reverse the endianness of the | |
604 | instructions back to little-endian, but leave the data as big-endian. | |
605 | ||
a05a5b64 | 606 | @cindex @code{-EL} command-line option, ARM |
252b5132 RH |
607 | @item -EL |
608 | This option specifies that the output generated by the assembler should | |
609 | be marked as being encoded for a little-endian processor. | |
adcf07e6 | 610 | |
a05a5b64 | 611 | @cindex @code{-k} command-line option, ARM |
252b5132 RH |
612 | @cindex PIC code generation for ARM |
613 | @item -k | |
a349d9dd PB |
614 | This option specifies that the output of the assembler should be marked |
615 | as position-independent code (PIC). | |
adcf07e6 | 616 | |
a05a5b64 | 617 | @cindex @code{--fix-v4bx} command-line option, ARM |
845b51d6 PB |
618 | @item --fix-v4bx |
619 | Allow @code{BX} instructions in ARMv4 code. This is intended for use with | |
620 | the linker option of the same name. | |
621 | ||
a05a5b64 | 622 | @cindex @code{-mwarn-deprecated} command-line option, ARM |
278df34e NS |
623 | @item -mwarn-deprecated |
624 | @itemx -mno-warn-deprecated | |
625 | Enable or disable warnings about using deprecated options or | |
626 | features. The default is to warn. | |
627 | ||
a05a5b64 | 628 | @cindex @code{-mccs} command-line option, ARM |
2e6976a8 DG |
629 | @item -mccs |
630 | Turns on CodeComposer Studio assembly syntax compatibility mode. | |
631 | ||
a05a5b64 | 632 | @cindex @code{-mwarn-syms} command-line option, ARM |
8b2d793c NC |
633 | @item -mwarn-syms |
634 | @itemx -mno-warn-syms | |
635 | Enable or disable warnings about symbols that match the names of ARM | |
636 | instructions. The default is to warn. | |
637 | ||
252b5132 RH |
638 | @end table |
639 | ||
640 | ||
641 | @node ARM Syntax | |
642 | @section Syntax | |
643 | @menu | |
cab7e4d9 | 644 | * ARM-Instruction-Set:: Instruction Set |
252b5132 RH |
645 | * ARM-Chars:: Special Characters |
646 | * ARM-Regs:: Register Names | |
b6895b4f | 647 | * ARM-Relocations:: Relocations |
99f1a7a7 | 648 | * ARM-Neon-Alignment:: NEON Alignment Specifiers |
252b5132 RH |
649 | @end menu |
650 | ||
cab7e4d9 NC |
651 | @node ARM-Instruction-Set |
652 | @subsection Instruction Set Syntax | |
653 | Two slightly different syntaxes are support for ARM and THUMB | |
654 | instructions. The default, @code{divided}, uses the old style where | |
655 | ARM and THUMB instructions had their own, separate syntaxes. The new, | |
656 | @code{unified} syntax, which can be selected via the @code{.syntax} | |
657 | directive, and has the following main features: | |
658 | ||
9e6f3811 AS |
659 | @itemize @bullet |
660 | @item | |
cab7e4d9 NC |
661 | Immediate operands do not require a @code{#} prefix. |
662 | ||
9e6f3811 | 663 | @item |
cab7e4d9 NC |
664 | The @code{IT} instruction may appear, and if it does it is validated |
665 | against subsequent conditional affixes. In ARM mode it does not | |
666 | generate machine code, in THUMB mode it does. | |
667 | ||
9e6f3811 | 668 | @item |
cab7e4d9 NC |
669 | For ARM instructions the conditional affixes always appear at the end |
670 | of the instruction. For THUMB instructions conditional affixes can be | |
671 | used, but only inside the scope of an @code{IT} instruction. | |
672 | ||
9e6f3811 | 673 | @item |
cab7e4d9 NC |
674 | All of the instructions new to the V6T2 architecture (and later) are |
675 | available. (Only a few such instructions can be written in the | |
676 | @code{divided} syntax). | |
677 | ||
9e6f3811 | 678 | @item |
cab7e4d9 NC |
679 | The @code{.N} and @code{.W} suffixes are recognized and honored. |
680 | ||
9e6f3811 | 681 | @item |
cab7e4d9 NC |
682 | All instructions set the flags if and only if they have an @code{s} |
683 | affix. | |
9e6f3811 | 684 | @end itemize |
cab7e4d9 | 685 | |
252b5132 RH |
686 | @node ARM-Chars |
687 | @subsection Special Characters | |
688 | ||
689 | @cindex line comment character, ARM | |
690 | @cindex ARM line comment character | |
7c31ae13 NC |
691 | The presence of a @samp{@@} anywhere on a line indicates the start of |
692 | a comment that extends to the end of that line. | |
693 | ||
694 | If a @samp{#} appears as the first character of a line then the whole | |
695 | line is treated as a comment, but in this case the line could also be | |
696 | a logical line number directive (@pxref{Comments}) or a preprocessor | |
697 | control command (@pxref{Preprocessing}). | |
550262c4 NC |
698 | |
699 | @cindex line separator, ARM | |
700 | @cindex statement separator, ARM | |
701 | @cindex ARM line separator | |
a349d9dd PB |
702 | The @samp{;} character can be used instead of a newline to separate |
703 | statements. | |
550262c4 NC |
704 | |
705 | @cindex immediate character, ARM | |
706 | @cindex ARM immediate character | |
707 | Either @samp{#} or @samp{$} can be used to indicate immediate operands. | |
252b5132 RH |
708 | |
709 | @cindex identifiers, ARM | |
710 | @cindex ARM identifiers | |
711 | *TODO* Explain about /data modifier on symbols. | |
712 | ||
713 | @node ARM-Regs | |
714 | @subsection Register Names | |
715 | ||
716 | @cindex ARM register names | |
717 | @cindex register names, ARM | |
718 | *TODO* Explain about ARM register naming, and the predefined names. | |
719 | ||
b6895b4f PB |
720 | @node ARM-Relocations |
721 | @subsection ARM relocation generation | |
722 | ||
723 | @cindex data relocations, ARM | |
724 | @cindex ARM data relocations | |
725 | Specific data relocations can be generated by putting the relocation name | |
726 | in parentheses after the symbol name. For example: | |
727 | ||
728 | @smallexample | |
729 | .word foo(TARGET1) | |
730 | @end smallexample | |
731 | ||
732 | This will generate an @samp{R_ARM_TARGET1} relocation against the symbol | |
733 | @var{foo}. | |
734 | The following relocations are supported: | |
735 | @code{GOT}, | |
736 | @code{GOTOFF}, | |
737 | @code{TARGET1}, | |
738 | @code{TARGET2}, | |
739 | @code{SBREL}, | |
740 | @code{TLSGD}, | |
741 | @code{TLSLDM}, | |
742 | @code{TLSLDO}, | |
0855e32b NS |
743 | @code{TLSDESC}, |
744 | @code{TLSCALL}, | |
b43420e6 NC |
745 | @code{GOTTPOFF}, |
746 | @code{GOT_PREL} | |
b6895b4f PB |
747 | and |
748 | @code{TPOFF}. | |
749 | ||
750 | For compatibility with older toolchains the assembler also accepts | |
3da1d841 NC |
751 | @code{(PLT)} after branch targets. On legacy targets this will |
752 | generate the deprecated @samp{R_ARM_PLT32} relocation. On EABI | |
753 | targets it will encode either the @samp{R_ARM_CALL} or | |
754 | @samp{R_ARM_JUMP24} relocation, as appropriate. | |
b6895b4f PB |
755 | |
756 | @cindex MOVW and MOVT relocations, ARM | |
757 | Relocations for @samp{MOVW} and @samp{MOVT} instructions can be generated | |
758 | by prefixing the value with @samp{#:lower16:} and @samp{#:upper16} | |
b45619c0 | 759 | respectively. For example to load the 32-bit address of foo into r0: |
252b5132 | 760 | |
b6895b4f PB |
761 | @smallexample |
762 | MOVW r0, #:lower16:foo | |
763 | MOVT r0, #:upper16:foo | |
764 | @end smallexample | |
252b5132 | 765 | |
72d98d16 MG |
766 | Relocations @samp{R_ARM_THM_ALU_ABS_G0_NC}, @samp{R_ARM_THM_ALU_ABS_G1_NC}, |
767 | @samp{R_ARM_THM_ALU_ABS_G2_NC} and @samp{R_ARM_THM_ALU_ABS_G3_NC} can be | |
768 | generated by prefixing the value with @samp{#:lower0_7:#}, | |
769 | @samp{#:lower8_15:#}, @samp{#:upper0_7:#} and @samp{#:upper8_15:#} | |
770 | respectively. For example to load the 32-bit address of foo into r0: | |
771 | ||
772 | @smallexample | |
773 | MOVS r0, #:upper8_15:#foo | |
774 | LSLS r0, r0, #8 | |
775 | ADDS r0, #:upper0_7:#foo | |
776 | LSLS r0, r0, #8 | |
777 | ADDS r0, #:lower8_15:#foo | |
778 | LSLS r0, r0, #8 | |
779 | ADDS r0, #:lower0_7:#foo | |
780 | @end smallexample | |
781 | ||
ba724cfc NC |
782 | @node ARM-Neon-Alignment |
783 | @subsection NEON Alignment Specifiers | |
784 | ||
785 | @cindex alignment for NEON instructions | |
786 | Some NEON load/store instructions allow an optional address | |
787 | alignment qualifier. | |
788 | The ARM documentation specifies that this is indicated by | |
789 | @samp{@@ @var{align}}. However GAS already interprets | |
790 | the @samp{@@} character as a "line comment" start, | |
791 | so @samp{: @var{align}} is used instead. For example: | |
792 | ||
793 | @smallexample | |
794 | vld1.8 @{q0@}, [r0, :128] | |
795 | @end smallexample | |
796 | ||
797 | @node ARM Floating Point | |
798 | @section Floating Point | |
799 | ||
800 | @cindex floating point, ARM (@sc{ieee}) | |
801 | @cindex ARM floating point (@sc{ieee}) | |
802 | The ARM family uses @sc{ieee} floating-point numbers. | |
803 | ||
252b5132 RH |
804 | @node ARM Directives |
805 | @section ARM Machine Directives | |
806 | ||
807 | @cindex machine directives, ARM | |
808 | @cindex ARM machine directives | |
809 | @table @code | |
810 | ||
4a6bc624 NS |
811 | @c AAAAAAAAAAAAAAAAAAAAAAAAA |
812 | ||
2b841ec2 | 813 | @ifclear ELF |
4a6bc624 NS |
814 | @cindex @code{.2byte} directive, ARM |
815 | @cindex @code{.4byte} directive, ARM | |
816 | @cindex @code{.8byte} directive, ARM | |
817 | @item .2byte @var{expression} [, @var{expression}]* | |
818 | @itemx .4byte @var{expression} [, @var{expression}]* | |
819 | @itemx .8byte @var{expression} [, @var{expression}]* | |
820 | These directives write 2, 4 or 8 byte values to the output section. | |
2b841ec2 | 821 | @end ifclear |
4a6bc624 NS |
822 | |
823 | @cindex @code{.align} directive, ARM | |
adcf07e6 NC |
824 | @item .align @var{expression} [, @var{expression}] |
825 | This is the generic @var{.align} directive. For the ARM however if the | |
826 | first argument is zero (ie no alignment is needed) the assembler will | |
827 | behave as if the argument had been 2 (ie pad to the next four byte | |
062b7c0c | 828 | boundary). This is for compatibility with ARM's own assembler. |
adcf07e6 | 829 | |
4a6bc624 NS |
830 | @cindex @code{.arch} directive, ARM |
831 | @item .arch @var{name} | |
832 | Select the target architecture. Valid values for @var{name} are the same as | |
54691107 TP |
833 | for the @option{-march} command-line option without the instruction set |
834 | extension. | |
252b5132 | 835 | |
34bca508 | 836 | Specifying @code{.arch} clears any previously selected architecture |
69133863 MGD |
837 | extensions. |
838 | ||
839 | @cindex @code{.arch_extension} directive, ARM | |
840 | @item .arch_extension @var{name} | |
34bca508 L |
841 | Add or remove an architecture extension to the target architecture. Valid |
842 | values for @var{name} are the same as those accepted as architectural | |
a05a5b64 | 843 | extensions by the @option{-mcpu} and @option{-march} command-line options. |
69133863 MGD |
844 | |
845 | @code{.arch_extension} may be used multiple times to add or remove extensions | |
846 | incrementally to the architecture being compiled for. | |
847 | ||
4a6bc624 NS |
848 | @cindex @code{.arm} directive, ARM |
849 | @item .arm | |
850 | This performs the same action as @var{.code 32}. | |
252b5132 | 851 | |
4a6bc624 | 852 | @c BBBBBBBBBBBBBBBBBBBBBBBBBB |
0bbf2aa4 | 853 | |
4a6bc624 NS |
854 | @cindex @code{.bss} directive, ARM |
855 | @item .bss | |
856 | This directive switches to the @code{.bss} section. | |
0bbf2aa4 | 857 | |
4a6bc624 NS |
858 | @c CCCCCCCCCCCCCCCCCCCCCCCCCC |
859 | ||
860 | @cindex @code{.cantunwind} directive, ARM | |
861 | @item .cantunwind | |
862 | Prevents unwinding through the current function. No personality routine | |
863 | or exception table data is required or permitted. | |
864 | ||
865 | @cindex @code{.code} directive, ARM | |
866 | @item .code @code{[16|32]} | |
867 | This directive selects the instruction set being generated. The value 16 | |
868 | selects Thumb, with the value 32 selecting ARM. | |
869 | ||
870 | @cindex @code{.cpu} directive, ARM | |
871 | @item .cpu @var{name} | |
872 | Select the target processor. Valid values for @var{name} are the same as | |
54691107 TP |
873 | for the @option{-mcpu} command-line option without the instruction set |
874 | extension. | |
4a6bc624 | 875 | |
34bca508 | 876 | Specifying @code{.cpu} clears any previously selected architecture |
69133863 MGD |
877 | extensions. |
878 | ||
4a6bc624 NS |
879 | @c DDDDDDDDDDDDDDDDDDDDDDDDDD |
880 | ||
881 | @cindex @code{.dn} and @code{.qn} directives, ARM | |
f467aa98 | 882 | @item @var{name} .dn @var{register name} [@var{.type}] [[@var{index}]] |
1f9bb1ca | 883 | @itemx @var{name} .qn @var{register name} [@var{.type}] [[@var{index}]] |
23753660 JB |
884 | |
885 | The @code{dn} and @code{qn} directives are used to create typed | |
886 | and/or indexed register aliases for use in Advanced SIMD Extension | |
887 | (Neon) instructions. The former should be used to create aliases | |
888 | of double-precision registers, and the latter to create aliases of | |
889 | quad-precision registers. | |
890 | ||
891 | If these directives are used to create typed aliases, those aliases can | |
892 | be used in Neon instructions instead of writing types after the mnemonic | |
893 | or after each operand. For example: | |
894 | ||
895 | @smallexample | |
896 | x .dn d2.f32 | |
897 | y .dn d3.f32 | |
898 | z .dn d4.f32[1] | |
899 | vmul x,y,z | |
900 | @end smallexample | |
901 | ||
902 | This is equivalent to writing the following: | |
903 | ||
904 | @smallexample | |
905 | vmul.f32 d2,d3,d4[1] | |
906 | @end smallexample | |
907 | ||
908 | Aliases created using @code{dn} or @code{qn} can be destroyed using | |
909 | @code{unreq}. | |
910 | ||
4a6bc624 | 911 | @c EEEEEEEEEEEEEEEEEEEEEEEEEE |
252b5132 | 912 | |
4a6bc624 NS |
913 | @cindex @code{.eabi_attribute} directive, ARM |
914 | @item .eabi_attribute @var{tag}, @var{value} | |
915 | Set the EABI object attribute @var{tag} to @var{value}. | |
252b5132 | 916 | |
4a6bc624 NS |
917 | The @var{tag} is either an attribute number, or one of the following: |
918 | @code{Tag_CPU_raw_name}, @code{Tag_CPU_name}, @code{Tag_CPU_arch}, | |
919 | @code{Tag_CPU_arch_profile}, @code{Tag_ARM_ISA_use}, | |
75375b3e | 920 | @code{Tag_THUMB_ISA_use}, @code{Tag_FP_arch}, @code{Tag_WMMX_arch}, |
a7ad558c | 921 | @code{Tag_Advanced_SIMD_arch}, @code{Tag_MVE_arch}, @code{Tag_PCS_config}, |
4a6bc624 NS |
922 | @code{Tag_ABI_PCS_R9_use}, @code{Tag_ABI_PCS_RW_data}, |
923 | @code{Tag_ABI_PCS_RO_data}, @code{Tag_ABI_PCS_GOT_use}, | |
924 | @code{Tag_ABI_PCS_wchar_t}, @code{Tag_ABI_FP_rounding}, | |
925 | @code{Tag_ABI_FP_denormal}, @code{Tag_ABI_FP_exceptions}, | |
926 | @code{Tag_ABI_FP_user_exceptions}, @code{Tag_ABI_FP_number_model}, | |
75375b3e | 927 | @code{Tag_ABI_align_needed}, @code{Tag_ABI_align_preserved}, |
4a6bc624 NS |
928 | @code{Tag_ABI_enum_size}, @code{Tag_ABI_HardFP_use}, |
929 | @code{Tag_ABI_VFP_args}, @code{Tag_ABI_WMMX_args}, | |
930 | @code{Tag_ABI_optimization_goals}, @code{Tag_ABI_FP_optimization_goals}, | |
931 | @code{Tag_compatibility}, @code{Tag_CPU_unaligned_access}, | |
75375b3e | 932 | @code{Tag_FP_HP_extension}, @code{Tag_ABI_FP_16bit_format}, |
cd21e546 | 933 | @code{Tag_MPextension_use}, @code{Tag_DIV_use}, |
4a6bc624 NS |
934 | @code{Tag_nodefaults}, @code{Tag_also_compatible_with}, |
935 | @code{Tag_conformance}, @code{Tag_T2EE_use}, | |
cd21e546 | 936 | @code{Tag_Virtualization_use} |
4a6bc624 NS |
937 | |
938 | The @var{value} is either a @code{number}, @code{"string"}, or | |
939 | @code{number, "string"} depending on the tag. | |
940 | ||
75375b3e | 941 | Note - the following legacy values are also accepted by @var{tag}: |
34bca508 | 942 | @code{Tag_VFP_arch}, @code{Tag_ABI_align8_needed}, |
75375b3e MGD |
943 | @code{Tag_ABI_align8_preserved}, @code{Tag_VFP_HP_extension}, |
944 | ||
4a6bc624 NS |
945 | @cindex @code{.even} directive, ARM |
946 | @item .even | |
947 | This directive aligns to an even-numbered address. | |
948 | ||
949 | @cindex @code{.extend} directive, ARM | |
950 | @cindex @code{.ldouble} directive, ARM | |
951 | @item .extend @var{expression} [, @var{expression}]* | |
952 | @itemx .ldouble @var{expression} [, @var{expression}]* | |
953 | These directives write 12byte long double floating-point values to the | |
954 | output section. These are not compatible with current ARM processors | |
955 | or ABIs. | |
956 | ||
957 | @c FFFFFFFFFFFFFFFFFFFFFFFFFF | |
958 | ||
5312fe52 BW |
959 | @cindex @code{.float16} directive, ARM |
960 | @item .float16 @var{value [,...,value_n]} | |
961 | Place the half precision floating point representation of one or more | |
962 | floating-point values into the current section. The exact format of the | |
963 | encoding is specified by @code{.float16_format}. If the format has not | |
964 | been explicitly set yet (either via the @code{.float16_format} directive or | |
965 | the command line option) then the IEEE 754-2008 format is used. | |
966 | ||
967 | @cindex @code{.float16_format} directive, ARM | |
968 | @item .float16_format @var{format} | |
969 | Set the format to use when encoding float16 values emitted by | |
970 | the @code{.float16} directive. | |
971 | Once the format has been set it cannot be changed. | |
972 | @code{format} should be one of the following: @code{ieee} (encode in | |
973 | the IEEE 754-2008 half precision format) or @code{alternative} (encode in | |
974 | the Arm alternative half precision format). | |
975 | ||
4a6bc624 NS |
976 | @anchor{arm_fnend} |
977 | @cindex @code{.fnend} directive, ARM | |
978 | @item .fnend | |
979 | Marks the end of a function with an unwind table entry. The unwind index | |
980 | table entry is created when this directive is processed. | |
252b5132 | 981 | |
4a6bc624 NS |
982 | If no personality routine has been specified then standard personality |
983 | routine 0 or 1 will be used, depending on the number of unwind opcodes | |
984 | required. | |
985 | ||
986 | @anchor{arm_fnstart} | |
987 | @cindex @code{.fnstart} directive, ARM | |
988 | @item .fnstart | |
989 | Marks the start of a function with an unwind table entry. | |
990 | ||
991 | @cindex @code{.force_thumb} directive, ARM | |
252b5132 RH |
992 | @item .force_thumb |
993 | This directive forces the selection of Thumb instructions, even if the | |
994 | target processor does not support those instructions | |
995 | ||
4a6bc624 NS |
996 | @cindex @code{.fpu} directive, ARM |
997 | @item .fpu @var{name} | |
998 | Select the floating-point unit to assemble for. Valid values for @var{name} | |
a05a5b64 | 999 | are the same as for the @option{-mfpu} command-line option. |
252b5132 | 1000 | |
4a6bc624 NS |
1001 | @c GGGGGGGGGGGGGGGGGGGGGGGGGG |
1002 | @c HHHHHHHHHHHHHHHHHHHHHHHHHH | |
e1da3f5b | 1003 | |
4a6bc624 NS |
1004 | @cindex @code{.handlerdata} directive, ARM |
1005 | @item .handlerdata | |
1006 | Marks the end of the current function, and the start of the exception table | |
1007 | entry for that function. Anything between this directive and the | |
1008 | @code{.fnend} directive will be added to the exception table entry. | |
1009 | ||
1010 | Must be preceded by a @code{.personality} or @code{.personalityindex} | |
1011 | directive. | |
1012 | ||
1013 | @c IIIIIIIIIIIIIIIIIIIIIIIIII | |
c921be7d NC |
1014 | |
1015 | @cindex @code{.inst} directive, ARM | |
1016 | @item .inst @var{opcode} [ , @dots{} ] | |
1f9bb1ca AS |
1017 | @itemx .inst.n @var{opcode} [ , @dots{} ] |
1018 | @itemx .inst.w @var{opcode} [ , @dots{} ] | |
c921be7d NC |
1019 | Generates the instruction corresponding to the numerical value @var{opcode}. |
1020 | @code{.inst.n} and @code{.inst.w} allow the Thumb instruction size to be | |
1021 | specified explicitly, overriding the normal encoding rules. | |
1022 | ||
4a6bc624 NS |
1023 | @c JJJJJJJJJJJJJJJJJJJJJJJJJJ |
1024 | @c KKKKKKKKKKKKKKKKKKKKKKKKKK | |
1025 | @c LLLLLLLLLLLLLLLLLLLLLLLLLL | |
1026 | ||
1027 | @item .ldouble @var{expression} [, @var{expression}]* | |
1028 | See @code{.extend}. | |
5395a469 | 1029 | |
252b5132 RH |
1030 | @cindex @code{.ltorg} directive, ARM |
1031 | @item .ltorg | |
1032 | This directive causes the current contents of the literal pool to be | |
1033 | dumped into the current section (which is assumed to be the .text | |
1034 | section) at the current location (aligned to a word boundary). | |
3d0c9500 NC |
1035 | @code{GAS} maintains a separate literal pool for each section and each |
1036 | sub-section. The @code{.ltorg} directive will only affect the literal | |
1037 | pool of the current section and sub-section. At the end of assembly | |
1038 | all remaining, un-empty literal pools will automatically be dumped. | |
1039 | ||
1040 | Note - older versions of @code{GAS} would dump the current literal | |
1041 | pool any time a section change occurred. This is no longer done, since | |
1042 | it prevents accurate control of the placement of literal pools. | |
252b5132 | 1043 | |
4a6bc624 | 1044 | @c MMMMMMMMMMMMMMMMMMMMMMMMMM |
252b5132 | 1045 | |
4a6bc624 NS |
1046 | @cindex @code{.movsp} directive, ARM |
1047 | @item .movsp @var{reg} [, #@var{offset}] | |
1048 | Tell the unwinder that @var{reg} contains an offset from the current | |
1049 | stack pointer. If @var{offset} is not specified then it is assumed to be | |
1050 | zero. | |
7ed4c4c5 | 1051 | |
4a6bc624 NS |
1052 | @c NNNNNNNNNNNNNNNNNNNNNNNNNN |
1053 | @c OOOOOOOOOOOOOOOOOOOOOOOOOO | |
7ed4c4c5 | 1054 | |
4a6bc624 NS |
1055 | @cindex @code{.object_arch} directive, ARM |
1056 | @item .object_arch @var{name} | |
1057 | Override the architecture recorded in the EABI object attribute section. | |
1058 | Valid values for @var{name} are the same as for the @code{.arch} directive. | |
1059 | Typically this is useful when code uses runtime detection of CPU features. | |
7ed4c4c5 | 1060 | |
4a6bc624 NS |
1061 | @c PPPPPPPPPPPPPPPPPPPPPPPPPP |
1062 | ||
1063 | @cindex @code{.packed} directive, ARM | |
1064 | @item .packed @var{expression} [, @var{expression}]* | |
1065 | This directive writes 12-byte packed floating-point values to the | |
1066 | output section. These are not compatible with current ARM processors | |
1067 | or ABIs. | |
1068 | ||
ea4cff4f | 1069 | @anchor{arm_pad} |
4a6bc624 NS |
1070 | @cindex @code{.pad} directive, ARM |
1071 | @item .pad #@var{count} | |
1072 | Generate unwinder annotations for a stack adjustment of @var{count} bytes. | |
1073 | A positive value indicates the function prologue allocated stack space by | |
1074 | decrementing the stack pointer. | |
7ed4c4c5 NC |
1075 | |
1076 | @cindex @code{.personality} directive, ARM | |
1077 | @item .personality @var{name} | |
1078 | Sets the personality routine for the current function to @var{name}. | |
1079 | ||
1080 | @cindex @code{.personalityindex} directive, ARM | |
1081 | @item .personalityindex @var{index} | |
1082 | Sets the personality routine for the current function to the EABI standard | |
1083 | routine number @var{index} | |
1084 | ||
4a6bc624 NS |
1085 | @cindex @code{.pool} directive, ARM |
1086 | @item .pool | |
1087 | This is a synonym for .ltorg. | |
7ed4c4c5 | 1088 | |
4a6bc624 NS |
1089 | @c QQQQQQQQQQQQQQQQQQQQQQQQQQ |
1090 | @c RRRRRRRRRRRRRRRRRRRRRRRRRR | |
1091 | ||
1092 | @cindex @code{.req} directive, ARM | |
1093 | @item @var{name} .req @var{register name} | |
1094 | This creates an alias for @var{register name} called @var{name}. For | |
1095 | example: | |
1096 | ||
1097 | @smallexample | |
1098 | foo .req r0 | |
1099 | @end smallexample | |
1100 | ||
1101 | @c SSSSSSSSSSSSSSSSSSSSSSSSSS | |
7ed4c4c5 | 1102 | |
7da4f750 | 1103 | @anchor{arm_save} |
7ed4c4c5 NC |
1104 | @cindex @code{.save} directive, ARM |
1105 | @item .save @var{reglist} | |
1106 | Generate unwinder annotations to restore the registers in @var{reglist}. | |
1107 | The format of @var{reglist} is the same as the corresponding store-multiple | |
1108 | instruction. | |
1109 | ||
1110 | @smallexample | |
1111 | @exdent @emph{core registers} | |
1112 | .save @{r4, r5, r6, lr@} | |
1113 | stmfd sp!, @{r4, r5, r6, lr@} | |
1114 | @exdent @emph{FPA registers} | |
1115 | .save f4, 2 | |
1116 | sfmfd f4, 2, [sp]! | |
1117 | @exdent @emph{VFP registers} | |
1118 | .save @{d8, d9, d10@} | |
fa073d69 | 1119 | fstmdx sp!, @{d8, d9, d10@} |
7ed4c4c5 NC |
1120 | @exdent @emph{iWMMXt registers} |
1121 | .save @{wr10, wr11@} | |
1122 | wstrd wr11, [sp, #-8]! | |
1123 | wstrd wr10, [sp, #-8]! | |
1124 | or | |
1125 | .save wr11 | |
1126 | wstrd wr11, [sp, #-8]! | |
1127 | .save wr10 | |
1128 | wstrd wr10, [sp, #-8]! | |
1129 | @end smallexample | |
1130 | ||
7da4f750 | 1131 | @anchor{arm_setfp} |
7ed4c4c5 NC |
1132 | @cindex @code{.setfp} directive, ARM |
1133 | @item .setfp @var{fpreg}, @var{spreg} [, #@var{offset}] | |
4a6bc624 | 1134 | Make all unwinder annotations relative to a frame pointer. Without this |
7ed4c4c5 NC |
1135 | the unwinder will use offsets from the stack pointer. |
1136 | ||
a5b82cbe | 1137 | The syntax of this directive is the same as the @code{add} or @code{mov} |
7ed4c4c5 NC |
1138 | instruction used to set the frame pointer. @var{spreg} must be either |
1139 | @code{sp} or mentioned in a previous @code{.movsp} directive. | |
1140 | ||
1141 | @smallexample | |
1142 | .movsp ip | |
1143 | mov ip, sp | |
1144 | @dots{} | |
1145 | .setfp fp, ip, #4 | |
a5b82cbe | 1146 | add fp, ip, #4 |
7ed4c4c5 NC |
1147 | @end smallexample |
1148 | ||
4a6bc624 NS |
1149 | @cindex @code{.secrel32} directive, ARM |
1150 | @item .secrel32 @var{expression} [, @var{expression}]* | |
1151 | This directive emits relocations that evaluate to the section-relative | |
1152 | offset of each expression's symbol. This directive is only supported | |
1153 | for PE targets. | |
1154 | ||
cab7e4d9 NC |
1155 | @cindex @code{.syntax} directive, ARM |
1156 | @item .syntax [@code{unified} | @code{divided}] | |
1157 | This directive sets the Instruction Set Syntax as described in the | |
1158 | @ref{ARM-Instruction-Set} section. | |
1159 | ||
4a6bc624 NS |
1160 | @c TTTTTTTTTTTTTTTTTTTTTTTTTT |
1161 | ||
1162 | @cindex @code{.thumb} directive, ARM | |
1163 | @item .thumb | |
1164 | This performs the same action as @var{.code 16}. | |
1165 | ||
1166 | @cindex @code{.thumb_func} directive, ARM | |
1167 | @item .thumb_func | |
1168 | This directive specifies that the following symbol is the name of a | |
1169 | Thumb encoded function. This information is necessary in order to allow | |
1170 | the assembler and linker to generate correct code for interworking | |
1171 | between Arm and Thumb instructions and should be used even if | |
1172 | interworking is not going to be performed. The presence of this | |
1173 | directive also implies @code{.thumb} | |
1174 | ||
33eaf5de | 1175 | This directive is not necessary when generating EABI objects. On these |
4a6bc624 NS |
1176 | targets the encoding is implicit when generating Thumb code. |
1177 | ||
1178 | @cindex @code{.thumb_set} directive, ARM | |
1179 | @item .thumb_set | |
1180 | This performs the equivalent of a @code{.set} directive in that it | |
1181 | creates a symbol which is an alias for another symbol (possibly not yet | |
1182 | defined). This directive also has the added property in that it marks | |
1183 | the aliased symbol as being a thumb function entry point, in the same | |
1184 | way that the @code{.thumb_func} directive does. | |
1185 | ||
0855e32b NS |
1186 | @cindex @code{.tlsdescseq} directive, ARM |
1187 | @item .tlsdescseq @var{tls-variable} | |
1188 | This directive is used to annotate parts of an inlined TLS descriptor | |
1189 | trampoline. Normally the trampoline is provided by the linker, and | |
1190 | this directive is not needed. | |
1191 | ||
4a6bc624 NS |
1192 | @c UUUUUUUUUUUUUUUUUUUUUUUUUU |
1193 | ||
1194 | @cindex @code{.unreq} directive, ARM | |
1195 | @item .unreq @var{alias-name} | |
1196 | This undefines a register alias which was previously defined using the | |
1197 | @code{req}, @code{dn} or @code{qn} directives. For example: | |
1198 | ||
1199 | @smallexample | |
1200 | foo .req r0 | |
1201 | .unreq foo | |
1202 | @end smallexample | |
1203 | ||
1204 | An error occurs if the name is undefined. Note - this pseudo op can | |
1205 | be used to delete builtin in register name aliases (eg 'r0'). This | |
1206 | should only be done if it is really necessary. | |
1207 | ||
7ed4c4c5 | 1208 | @cindex @code{.unwind_raw} directive, ARM |
4a6bc624 | 1209 | @item .unwind_raw @var{offset}, @var{byte1}, @dots{} |
33eaf5de | 1210 | Insert one of more arbitrary unwind opcode bytes, which are known to adjust |
7ed4c4c5 NC |
1211 | the stack pointer by @var{offset} bytes. |
1212 | ||
1213 | For example @code{.unwind_raw 4, 0xb1, 0x01} is equivalent to | |
1214 | @code{.save @{r0@}} | |
1215 | ||
4a6bc624 | 1216 | @c VVVVVVVVVVVVVVVVVVVVVVVVVV |
ee065d83 | 1217 | |
4a6bc624 NS |
1218 | @cindex @code{.vsave} directive, ARM |
1219 | @item .vsave @var{vfp-reglist} | |
1220 | Generate unwinder annotations to restore the VFP registers in @var{vfp-reglist} | |
1221 | using FLDMD. Also works for VFPv3 registers | |
1222 | that are to be restored using VLDM. | |
1223 | The format of @var{vfp-reglist} is the same as the corresponding store-multiple | |
1224 | instruction. | |
ee065d83 | 1225 | |
4a6bc624 NS |
1226 | @smallexample |
1227 | @exdent @emph{VFP registers} | |
1228 | .vsave @{d8, d9, d10@} | |
1229 | fstmdd sp!, @{d8, d9, d10@} | |
1230 | @exdent @emph{VFPv3 registers} | |
1231 | .vsave @{d15, d16, d17@} | |
1232 | vstm sp!, @{d15, d16, d17@} | |
1233 | @end smallexample | |
e04befd0 | 1234 | |
4a6bc624 NS |
1235 | Since FLDMX and FSTMX are now deprecated, this directive should be |
1236 | used in favour of @code{.save} for saving VFP registers for ARMv6 and above. | |
e04befd0 | 1237 | |
4a6bc624 NS |
1238 | @c WWWWWWWWWWWWWWWWWWWWWWWWWW |
1239 | @c XXXXXXXXXXXXXXXXXXXXXXXXXX | |
1240 | @c YYYYYYYYYYYYYYYYYYYYYYYYYY | |
1241 | @c ZZZZZZZZZZZZZZZZZZZZZZZZZZ | |
ee065d83 | 1242 | |
252b5132 RH |
1243 | @end table |
1244 | ||
1245 | @node ARM Opcodes | |
1246 | @section Opcodes | |
1247 | ||
1248 | @cindex ARM opcodes | |
1249 | @cindex opcodes for ARM | |
49a5575c NC |
1250 | @code{@value{AS}} implements all the standard ARM opcodes. It also |
1251 | implements several pseudo opcodes, including several synthetic load | |
34bca508 | 1252 | instructions. |
252b5132 | 1253 | |
49a5575c NC |
1254 | @table @code |
1255 | ||
1256 | @cindex @code{NOP} pseudo op, ARM | |
1257 | @item NOP | |
1258 | @smallexample | |
1259 | nop | |
1260 | @end smallexample | |
252b5132 | 1261 | |
49a5575c NC |
1262 | This pseudo op will always evaluate to a legal ARM instruction that does |
1263 | nothing. Currently it will evaluate to MOV r0, r0. | |
252b5132 | 1264 | |
49a5575c | 1265 | @cindex @code{LDR reg,=<label>} pseudo op, ARM |
34bca508 | 1266 | @item LDR |
252b5132 RH |
1267 | @smallexample |
1268 | ldr <register> , = <expression> | |
1269 | @end smallexample | |
1270 | ||
1271 | If expression evaluates to a numeric constant then a MOV or MVN | |
1272 | instruction will be used in place of the LDR instruction, if the | |
1273 | constant can be generated by either of these instructions. Otherwise | |
1274 | the constant will be placed into the nearest literal pool (if it not | |
1275 | already there) and a PC relative LDR instruction will be generated. | |
1276 | ||
49a5575c NC |
1277 | @cindex @code{ADR reg,<label>} pseudo op, ARM |
1278 | @item ADR | |
1279 | @smallexample | |
1280 | adr <register> <label> | |
1281 | @end smallexample | |
1282 | ||
1283 | This instruction will load the address of @var{label} into the indicated | |
1284 | register. The instruction will evaluate to a PC relative ADD or SUB | |
1285 | instruction depending upon where the label is located. If the label is | |
1286 | out of range, or if it is not defined in the same file (and section) as | |
1287 | the ADR instruction, then an error will be generated. This instruction | |
1288 | will not make use of the literal pool. | |
1289 | ||
fc6141f0 NC |
1290 | If @var{label} is a thumb function symbol, and thumb interworking has |
1291 | been enabled via the @option{-mthumb-interwork} option then the bottom | |
1292 | bit of the value stored into @var{register} will be set. This allows | |
1293 | the following sequence to work as expected: | |
1294 | ||
1295 | @smallexample | |
1296 | adr r0, thumb_function | |
1297 | blx r0 | |
1298 | @end smallexample | |
1299 | ||
49a5575c | 1300 | @cindex @code{ADRL reg,<label>} pseudo op, ARM |
34bca508 | 1301 | @item ADRL |
49a5575c NC |
1302 | @smallexample |
1303 | adrl <register> <label> | |
1304 | @end smallexample | |
1305 | ||
1306 | This instruction will load the address of @var{label} into the indicated | |
a349d9dd | 1307 | register. The instruction will evaluate to one or two PC relative ADD |
49a5575c NC |
1308 | or SUB instructions depending upon where the label is located. If a |
1309 | second instruction is not needed a NOP instruction will be generated in | |
1310 | its place, so that this instruction is always 8 bytes long. | |
1311 | ||
1312 | If the label is out of range, or if it is not defined in the same file | |
1313 | (and section) as the ADRL instruction, then an error will be generated. | |
1314 | This instruction will not make use of the literal pool. | |
1315 | ||
fc6141f0 NC |
1316 | If @var{label} is a thumb function symbol, and thumb interworking has |
1317 | been enabled via the @option{-mthumb-interwork} option then the bottom | |
1318 | bit of the value stored into @var{register} will be set. | |
1319 | ||
49a5575c NC |
1320 | @end table |
1321 | ||
252b5132 RH |
1322 | For information on the ARM or Thumb instruction sets, see @cite{ARM |
1323 | Software Development Toolkit Reference Manual}, Advanced RISC Machines | |
1324 | Ltd. | |
1325 | ||
6057a28f NC |
1326 | @node ARM Mapping Symbols |
1327 | @section Mapping Symbols | |
1328 | ||
1329 | The ARM ELF specification requires that special symbols be inserted | |
1330 | into object files to mark certain features: | |
1331 | ||
1332 | @table @code | |
1333 | ||
1334 | @cindex @code{$a} | |
1335 | @item $a | |
1336 | At the start of a region of code containing ARM instructions. | |
1337 | ||
1338 | @cindex @code{$t} | |
1339 | @item $t | |
1340 | At the start of a region of code containing THUMB instructions. | |
1341 | ||
1342 | @cindex @code{$d} | |
1343 | @item $d | |
1344 | At the start of a region of data. | |
1345 | ||
1346 | @end table | |
1347 | ||
1348 | The assembler will automatically insert these symbols for you - there | |
1349 | is no need to code them yourself. Support for tagging symbols ($b, | |
1350 | $f, $p and $m) which is also mentioned in the current ARM ELF | |
1351 | specification is not implemented. This is because they have been | |
1352 | dropped from the new EABI and so tools cannot rely upon their | |
1353 | presence. | |
1354 | ||
7da4f750 MM |
1355 | @node ARM Unwinding Tutorial |
1356 | @section Unwinding | |
1357 | ||
1358 | The ABI for the ARM Architecture specifies a standard format for | |
1359 | exception unwind information. This information is used when an | |
1360 | exception is thrown to determine where control should be transferred. | |
1361 | In particular, the unwind information is used to determine which | |
1362 | function called the function that threw the exception, and which | |
1363 | function called that one, and so forth. This information is also used | |
1364 | to restore the values of callee-saved registers in the function | |
1365 | catching the exception. | |
1366 | ||
1367 | If you are writing functions in assembly code, and those functions | |
1368 | call other functions that throw exceptions, you must use assembly | |
1369 | pseudo ops to ensure that appropriate exception unwind information is | |
1370 | generated. Otherwise, if one of the functions called by your assembly | |
1371 | code throws an exception, the run-time library will be unable to | |
1372 | unwind the stack through your assembly code and your program will not | |
1373 | behave correctly. | |
1374 | ||
1375 | To illustrate the use of these pseudo ops, we will examine the code | |
1376 | that G++ generates for the following C++ input: | |
1377 | ||
1378 | @verbatim | |
1379 | void callee (int *); | |
1380 | ||
34bca508 L |
1381 | int |
1382 | caller () | |
7da4f750 MM |
1383 | { |
1384 | int i; | |
1385 | callee (&i); | |
34bca508 | 1386 | return i; |
7da4f750 MM |
1387 | } |
1388 | @end verbatim | |
1389 | ||
1390 | This example does not show how to throw or catch an exception from | |
1391 | assembly code. That is a much more complex operation and should | |
1392 | always be done in a high-level language, such as C++, that directly | |
1393 | supports exceptions. | |
1394 | ||
1395 | The code generated by one particular version of G++ when compiling the | |
1396 | example above is: | |
1397 | ||
1398 | @verbatim | |
1399 | _Z6callerv: | |
1400 | .fnstart | |
1401 | .LFB2: | |
1402 | @ Function supports interworking. | |
1403 | @ args = 0, pretend = 0, frame = 8 | |
1404 | @ frame_needed = 1, uses_anonymous_args = 0 | |
1405 | stmfd sp!, {fp, lr} | |
1406 | .save {fp, lr} | |
1407 | .LCFI0: | |
1408 | .setfp fp, sp, #4 | |
1409 | add fp, sp, #4 | |
1410 | .LCFI1: | |
1411 | .pad #8 | |
1412 | sub sp, sp, #8 | |
1413 | .LCFI2: | |
1414 | sub r3, fp, #8 | |
1415 | mov r0, r3 | |
1416 | bl _Z6calleePi | |
1417 | ldr r3, [fp, #-8] | |
1418 | mov r0, r3 | |
1419 | sub sp, fp, #4 | |
1420 | ldmfd sp!, {fp, lr} | |
1421 | bx lr | |
1422 | .LFE2: | |
1423 | .fnend | |
1424 | @end verbatim | |
1425 | ||
1426 | Of course, the sequence of instructions varies based on the options | |
1427 | you pass to GCC and on the version of GCC in use. The exact | |
1428 | instructions are not important since we are focusing on the pseudo ops | |
1429 | that are used to generate unwind information. | |
1430 | ||
1431 | An important assumption made by the unwinder is that the stack frame | |
1432 | does not change during the body of the function. In particular, since | |
1433 | we assume that the assembly code does not itself throw an exception, | |
1434 | the only point where an exception can be thrown is from a call, such | |
1435 | as the @code{bl} instruction above. At each call site, the same saved | |
1436 | registers (including @code{lr}, which indicates the return address) | |
1437 | must be located in the same locations relative to the frame pointer. | |
1438 | ||
1439 | The @code{.fnstart} (@pxref{arm_fnstart,,.fnstart pseudo op}) pseudo | |
1440 | op appears immediately before the first instruction of the function | |
1441 | while the @code{.fnend} (@pxref{arm_fnend,,.fnend pseudo op}) pseudo | |
1442 | op appears immediately after the last instruction of the function. | |
34bca508 | 1443 | These pseudo ops specify the range of the function. |
7da4f750 MM |
1444 | |
1445 | Only the order of the other pseudos ops (e.g., @code{.setfp} or | |
1446 | @code{.pad}) matters; their exact locations are irrelevant. In the | |
1447 | example above, the compiler emits the pseudo ops with particular | |
1448 | instructions. That makes it easier to understand the code, but it is | |
1449 | not required for correctness. It would work just as well to emit all | |
1450 | of the pseudo ops other than @code{.fnend} in the same order, but | |
1451 | immediately after @code{.fnstart}. | |
1452 | ||
1453 | The @code{.save} (@pxref{arm_save,,.save pseudo op}) pseudo op | |
1454 | indicates registers that have been saved to the stack so that they can | |
1455 | be restored before the function returns. The argument to the | |
1456 | @code{.save} pseudo op is a list of registers to save. If a register | |
1457 | is ``callee-saved'' (as specified by the ABI) and is modified by the | |
1458 | function you are writing, then your code must save the value before it | |
1459 | is modified and restore the original value before the function | |
1460 | returns. If an exception is thrown, the run-time library restores the | |
1461 | values of these registers from their locations on the stack before | |
1462 | returning control to the exception handler. (Of course, if an | |
1463 | exception is not thrown, the function that contains the @code{.save} | |
1464 | pseudo op restores these registers in the function epilogue, as is | |
1465 | done with the @code{ldmfd} instruction above.) | |
1466 | ||
1467 | You do not have to save callee-saved registers at the very beginning | |
1468 | of the function and you do not need to use the @code{.save} pseudo op | |
1469 | immediately following the point at which the registers are saved. | |
1470 | However, if you modify a callee-saved register, you must save it on | |
1471 | the stack before modifying it and before calling any functions which | |
1472 | might throw an exception. And, you must use the @code{.save} pseudo | |
1473 | op to indicate that you have done so. | |
1474 | ||
1475 | The @code{.pad} (@pxref{arm_pad,,.pad}) pseudo op indicates a | |
1476 | modification of the stack pointer that does not save any registers. | |
1477 | The argument is the number of bytes (in decimal) that are subtracted | |
1478 | from the stack pointer. (On ARM CPUs, the stack grows downwards, so | |
1479 | subtracting from the stack pointer increases the size of the stack.) | |
1480 | ||
1481 | The @code{.setfp} (@pxref{arm_setfp,,.setfp pseudo op}) pseudo op | |
1482 | indicates the register that contains the frame pointer. The first | |
1483 | argument is the register that is set, which is typically @code{fp}. | |
1484 | The second argument indicates the register from which the frame | |
1485 | pointer takes its value. The third argument, if present, is the value | |
1486 | (in decimal) added to the register specified by the second argument to | |
1487 | compute the value of the frame pointer. You should not modify the | |
1488 | frame pointer in the body of the function. | |
1489 | ||
1490 | If you do not use a frame pointer, then you should not use the | |
1491 | @code{.setfp} pseudo op. If you do not use a frame pointer, then you | |
1492 | should avoid modifying the stack pointer outside of the function | |
1493 | prologue. Otherwise, the run-time library will be unable to find | |
1494 | saved registers when it is unwinding the stack. | |
1495 | ||
1496 | The pseudo ops described above are sufficient for writing assembly | |
1497 | code that calls functions which may throw exceptions. If you need to | |
1498 | know more about the object-file format used to represent unwind | |
1499 | information, you may consult the @cite{Exception Handling ABI for the | |
1500 | ARM Architecture} available from @uref{http://infocenter.arm.com}. | |
91f68a68 | 1501 |