1 .\" Copyright (c) 2016-2017 The FreeBSD Foundation. All rights reserved.
3 .\" This documentation was created by Ed Maste under sponsorship of
4 .\" The FreeBSD Foundation.
6 .\" Redistribution and use in source and binary forms, with or without
7 .\" modification, are permitted provided that the following conditions
9 .\" 1. Redistributions of source code must retain the above copyright
10 .\" notice, this list of conditions and the following disclaimer.
11 .\" 2. Redistributions in binary form must reproduce the above copyright
12 .\" notice, this list of conditions and the following disclaimer in the
13 .\" documentation and/or other materials provided with the distribution.
15 .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND
16 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE
19 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 .Nd Architecture-specific details
36 Differences between CPU architectures and platforms supported by
39 This document is a quick reference of key ABI details of
42 For full details consult the processor-specific ABI supplement
45 If not explicitly mentioned, sizes are in bytes.
46 The architecture details in this document apply to
48 and later, unless otherwise noted.
51 uses a flat address space.
57 and pointers all have the same representation.
59 In order to maximize compatibility with future pointer integrity mechanisms,
60 manipulations of pointers as integers should be performed via
71 On some architectures, e.g.,
76 the kernel uses a separate address space.
77 On other architectures, kernel and a user mode process share a
79 The kernel is located at the highest addresses.
81 On each architecture, the main user mode thread's stack starts near
82 the highest user address and grows down.
85 architecture support varies by release.
86 This table shows the first
88 release to support each architecture, and, for discontinued
89 architectures, the final release.
91 .Bl -column -offset indent "Sy Architecture" "Sy Initial Release" "Sy Final Release"
92 .It Sy Architecture Ta Sy Initial Release Ta Sy Final Release
94 .It alpha Ta 3.2 Ta 6.4
96 .It arm Ta 6.0 Ta 12.x
97 .It armeb Ta 8.0 Ta 11.x
100 .It ia64 Ta 5.0 Ta 10.4
109 .It mips64elhf Ta 12.0
111 .It pc98 Ta 2.2 Ta 11.x
113 .It powerpcspe Ta 12.0
116 .It riscv64sf Ta 12.0
117 .It sparc64 Ta 5.0 Ta 12.x
122 architectures use some variant of the ELF (see
124 .Sy Application Binary Interface
125 (ABI) for the machine processor.
126 All supported ABIs can be divided into two groups:
127 .Bl -tag -width "Dv ILP32"
132 types machine representations all have 4-byte size.
135 type machine representation uses 4 bytes,
145 symbol when compiling for an
149 Some machines support more than one
152 Typically these are 64-bit machines, where the
155 execution environment is accompanied by the
158 environment, which was the historical 32-bit predecessor for 64-bit evolution.
160 .Bl -column -offset indent "Dv powerpc64" "Sy ILP32 counterpart"
161 .It Sy LP64 Ta Sy ILP32 counterpart
162 .It Dv amd64 Ta Dv i386
163 .It Dv powerpc64 Ta Dv powerpc
164 .It Dv mips64* Ta Dv mips*
165 .It Dv aarch64 Ta Dv armv6/armv7
169 will support execution of
173 binaries if the CPU implements
175 execution state, however
177 binaries aren't supported.
179 On all supported architectures:
180 .Bl -column -offset -indent "long long" "Size"
181 .It Sy Type Ta Sy Size
184 .It long Ta sizeof(void*)
190 Integers are represented in two's complement.
191 Alignment of integer and pointer types is natural, that is,
192 the address of the variable must be congruent to zero modulo the type size.
193 Most ILP32 ABIs, except
195 require only 4-byte alignment for 64-bit integers.
197 Machine-dependent type sizes:
198 .Bl -column -offset indent "Sy Architecture" "Sy void *" "Sy long double" "Sy time_t"
199 .It Sy Architecture Ta Sy void * Ta Sy long double Ta Sy time_t
200 .It aarch64 Ta 8 Ta 16 Ta 8
201 .It amd64 Ta 8 Ta 16 Ta 8
202 .It arm Ta 4 Ta 8 Ta 8
203 .It armv6 Ta 4 Ta 8 Ta 8
204 .It i386 Ta 4 Ta 12 Ta 4
205 .It mips Ta 4 Ta 8 Ta 8
206 .It mipsel Ta 4 Ta 8 Ta 8
207 .It mipselhf Ta 4 Ta 8 Ta 8
208 .It mipshf Ta 4 Ta 8 Ta 8
209 .It mipsn32 Ta 4 Ta 8 Ta 8
210 .It mips64 Ta 8 Ta 8 Ta 8
211 .It mips64el Ta 8 Ta 8 Ta 8
212 .It mips64elhf Ta 8 Ta 8 Ta 8
213 .It mips64hf Ta 8 Ta 8 Ta 8
214 .It powerpc Ta 4 Ta 8 Ta 8
215 .It powerpcspe Ta 4 Ta 8 Ta 8
216 .It powerpc64 Ta 8 Ta 8 Ta 8
217 .It riscv64 Ta 8 Ta 16 Ta 8
218 .It riscv64sf Ta 8 Ta 16 Ta 8
219 .It sparc64 Ta 8 Ta 16 Ta 8
223 is 8 bytes on all supported architectures except i386.
224 .Ss Endianness and Char Signedness
225 .Bl -column -offset indent "Sy Architecture" "Sy Endianness" "Sy char Signedness"
226 .It Sy Architecture Ta Sy Endianness Ta Sy char Signedness
227 .It aarch64 Ta little Ta unsigned
228 .It amd64 Ta little Ta signed
229 .It arm Ta little Ta unsigned
230 .It armv6 Ta little Ta unsigned
231 .It armv7 Ta little Ta unsigned
232 .It i386 Ta little Ta signed
233 .It mips Ta big Ta signed
234 .It mipsel Ta little Ta signed
235 .It mipselhf Ta little Ta signed
236 .It mipshf Ta big Ta signed
237 .It mipsn32 Ta big Ta signed
238 .It mips64 Ta big Ta signed
239 .It mips64el Ta little Ta signed
240 .It mips64elhf Ta little Ta signed
241 .It mips64hf Ta big Ta signed
242 .It powerpc Ta big Ta unsigned
243 .It powerpcspe Ta big Ta unsigned
244 .It powerpc64 Ta big Ta unsigned
245 .It riscv64 Ta little Ta signed
246 .It riscv64sf Ta little Ta signed
247 .It sparc64 Ta big Ta signed
250 .Bl -column -offset indent "Sy Architecture" "Sy Page Sizes"
251 .It Sy Architecture Ta Sy Page Sizes
252 .It aarch64 Ta 4K, 2M, 1G
253 .It amd64 Ta 4K, 2M, 1G
257 .It i386 Ta 4K, 2M (PAE), 4M
275 .Bl -column -offset indent "Sy Architecture" "Sy float, double" "Sy long double"
276 .It Sy Architecture Ta Sy float, double Ta Sy long double
277 .It aarch64 Ta hard Ta soft, quad precision
278 .It amd64 Ta hard Ta hard, 80 bit
279 .It arm Ta soft Ta soft, double precision
280 .It armv6 Ta hard Ta hard, double precision
281 .It armv7 Ta hard Ta hard, double precision
282 .It i386 Ta hard Ta hard, 80 bit
283 .It mips Ta soft Ta identical to double
284 .It mipsel Ta soft Ta identical to double
285 .It mipselhf Ta hard Ta identical to double
286 .It mipshf Ta hard Ta identical to double
287 .It mipsn32 Ta soft Ta identical to double
288 .It mips64 Ta soft Ta identical to double
289 .It mips64el Ta soft Ta identical to double
290 .It mips64elhf Ta hard Ta identical to double
291 .It mips64hf Ta hard Ta identical to double
292 .It powerpc Ta hard Ta hard, double precision
293 .It powerpcspe Ta hard Ta hard, double precision
294 .It powerpc64 Ta hard Ta hard, double precision
295 .It riscv64 Ta hard Ta hard, double precision
296 .It riscv64sf Ta soft Ta soft, double precision
297 .It sparc64 Ta hard Ta hard, quad precision
299 .Ss Default Tool Chain
300 .Fx uses a variety of tool chain components for the supported CPU
305 provided by the base system,
310 or an external toolchain compiler and linker provided by a port or package.
311 This table shows the default tool chain for each architecture.
312 .Bl -column -offset indent "Sy Architecture" "Sy Compiler" "Sy Linker"
313 .It Sy Architecture Ta Sy Compiler Ta Sy Linker
314 .It aarch64 Ta Clang Ta lld
315 .It amd64 Ta Clang Ta lld
316 .It arm Ta Clang Ta GNU ld 2.17.50
317 .It armv6 Ta Clang Ta lld
318 .It armv7 Ta Clang Ta lld
319 .It i386 Ta Clang Ta lld
320 .It mips Ta GCC 4.2.1 Ta GNU ld 2.17.50
321 .It mipsel Ta GCC 4.2.1 Ta GNU ld 2.17.50
322 .It mipselhf Ta GCC 4.2.1 Ta GNU ld 2.17.50
323 .It mipshf Ta GCC 4.2.1 Ta GNU ld 2.17.50
324 .It mipsn32 Ta GCC 4.2.1 Ta GNU ld 2.17.50
325 .It mips64 Ta GCC 4.2.1 Ta GNU ld 2.17.50
326 .It mips64el Ta GCC 4.2.1 Ta GNU ld 2.17.50
327 .It mips64elhf Ta GCC 4.2.1 Ta GNU ld 2.17.50
328 .It mips64hf Ta GCC 4.2.1 Ta GNU ld 2.17.50
329 .It powerpc Ta Clang Ta GNU ld 2.17.50
330 .It powerpcspe Ta Clang Ta GNU ld 2.17.50
331 .It powerpc64 Ta Clang Ta lld
332 .It riscv64 Ta GCC(1) Ta GNU ld(1)
333 .It riscv64sf Ta GCC(1) Ta GNU ld(1)
334 .It sparc64 Ta GCC 4.2.1 Ta GNU ld 2.17.50
337 (1) External toolchain provided by ports/packages.
339 Note that GCC 4.2.1 is deprecated, and scheduled for removal on 2020-03-31.
340 Any CPU architectures not migrated by then
341 (to either base system Clang or external toolchain)
342 may be removed from the tree after that date.
343 Unless the make variable
344 .Dv MAKE_OBSOLETE_GCC
345 is defined, make universe will not build mips or sparc64
346 architectures unless the xtoolchain binaries have been installed for
348 .Ss MACHINE_ARCH vs MACHINE_CPUARCH vs MACHINE
350 should be preferred in Makefiles when the generic
351 architecture is being tested.
353 should be preferred when there is something specific to a particular type of
354 architecture where there is a choice of many, or could be a choice of many.
357 when referring to the kernel, interfaces dependent on a specific type of kernel
358 or similar things like boot sequences.
359 .Bl -column -offset indent "Dv MACHINE" "Dv MACHINE_CPUARCH" "Dv MACHINE_ARCH"
360 .It Dv MACHINE Ta Dv MACHINE_CPUARCH Ta Dv MACHINE_ARCH
361 .It arm64 Ta aarch64 Ta aarch64
362 .It amd64 Ta amd64 Ta amd64
363 .It arm Ta arm Ta arm, armv6, armv7
364 .It i386 Ta i386 Ta i386
365 .It mips Ta mips Ta mips, mipsel, mips64, mips64el, mipshf, mipselhf, mips64elhf, mipsn32
366 .It powerpc Ta powerpc Ta powerpc, powerpcspe, powerpc64
367 .It riscv Ta riscv Ta riscv64, riscv64sf
368 .It sparc64 Ta sparc64 Ta sparc64
370 .Ss Predefined Macros
371 The compiler provides a number of predefined macros.
372 Some of these provide architecture-specific details and are explained below.
373 Other macros, including those required by the language standard, are not
376 The full set of predefined macros can be obtained with this command:
377 .Bd -literal -offset indent
378 cc -x c -dM -E /dev/null
381 Common type size and endianness macros:
382 .Bl -column -offset indent "BYTE_ORDER" "Sy Meaning"
383 .It Sy Macro Ta Sy Meaning
384 .It Dv __LP64__ Ta 64-bit (8-byte) long and pointer, 32-bit (4-byte) int
385 .It Dv __ILP32__ Ta 32-bit (4-byte) int, long and pointer
386 .It Dv BYTE_ORDER Ta Either Dv BIG_ENDIAN or Dv LITTLE_ENDIAN .
392 Architecture-specific macros:
393 .Bl -column -offset indent "Sy Architecture" "Sy Predefined macros"
394 .It Sy Architecture Ta Sy Predefined macros
395 .It aarch64 Ta Dv __aarch64__
396 .It amd64 Ta Dv __amd64__, Dv __x86_64__
397 .It arm Ta Dv __arm__
398 .It armv6 Ta Dv __arm__, Dv __ARM_ARCH >= 6
399 .It armv7 Ta Dv __arm__, Dv __ARM_ARCH >= 7
400 .It i386 Ta Dv __i386__
401 .It mips Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_o32
402 .It mipsel Ta Dv __mips__, Dv __mips_o32
403 .It mipselhf Ta Dv __mips__, Dv __mips_o32
404 .It mipshf Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_o32
405 .It mipsn32 Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n32
406 .It mips64 Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n64
407 .It mips64el Ta Dv __mips__, Dv __mips_n64
408 .It mips64elhf Ta Dv __mips__, Dv __mips_n64
409 .It mips64hf Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n64
410 .It powerpc Ta Dv __powerpc__
411 .It powerpcspe Ta Dv __powerpc__, Dv __SPE__
412 .It powerpc64 Ta Dv __powerpc__, Dv __powerpc64__
413 .It riscv64 Ta Dv __riscv, Dv __riscv_xlen == 64
414 .It riscv64sf Ta Dv __riscv, Dv __riscv_xlen == 64
415 .It sparc64 Ta Dv __sparc64__
418 Compilers may define additional variants of architecture-specific macros.
419 The macros above are preferred for use in
421 .Ss Important Xr make 1 variables
422 Most of the externally settable variables are defined in the
425 These variables are not otherwise documented and are used extensively
427 .Bl -tag -width "MACHINE_CPUARCH"
429 Represent the hardware platform.
430 This is the same as the native platform's
434 It defines both the userland / kernel interface, as well as the
435 bootloader / kernel interface.
436 It should only be used in these contexts.
437 Each CPU architecture may have multiple hardware platforms it supports
441 It is used to collect together all the files from
444 It is often the same as
446 just as one CPU architecture can be implemented by many different
447 hardware platforms, one hardware platform may support multiple CPU
448 architecture family members, though with different binaries.
451 of i386 supported the IBM-AT hardware platform while the
453 of pc98 supported the Japanese company NEC's PC-9801 and PC-9821
455 Both of these hardware platforms supported only the
457 of i386 where they shared a common ABI, except for certain kernel /
458 userland interfaces relating to underlying hardware platform
459 differences in bus architecture, device enumeration and boot interface.
462 should only be used in src/sys and src/stand or in system imagers or
465 Represents the CPU processor architecture.
466 This is the same as the native platforms
470 It defines the CPU instruction family supported.
471 It may also encode a variation in the byte ordering of multi-byte
473 It may also encode a variation in the size of the integer or pointer.
474 It may also encode a ISA revision.
475 It may also encode hard versus soft floating point ABI and usage.
476 It may also encode a variant ABI when the other factors do not
477 uniquely define the ABI (e.g., MIPS' n32 ABI).
480 defines the ABI used by the system.
481 For example, the MIPS CPU processor family supports 9 different
482 combinations encoding pointer size, endian and hard versus soft float (for
483 8 combinations) as well as N32 (which only ever had one variation of
485 Generally, the plain CPU name specifies the most common (or at least
486 first) variant of the CPU.
487 This is why mips and mips64 imply 'big endian' while 'arm' and 'armv7'
489 If we ever were to support the so-called x32 ABI (using 32-bit
490 pointers on the amd64 architecture), it would most likely be encoded
492 It is unfortunate that amd64 specifies the 64-bit evolution of the x86
493 platform (it matches the 'first rule') as everybody else uses x86_64.
494 There is no standard name for the processor: each OS selects its own
496 .It Dv MACHINE_CPUARCH
497 Represents the source location for a given
499 It is generally the common prefix for all the MACHINE_ARCH that
500 share the same implementation, though 'riscv' breaks this rule.
503 is defined to be mips for all the flavors of mips that we support
504 since we support them all with a shared set of sources.
505 While amd64 and i386 are closely related, MACHINE_CPUARCH is not x86
507 The FreeBSD source base supports amd64 and i386 with two
508 distinct source bases living in subdirectories named amd64 and i386
509 (though behind the scenes there's some sharing that fits into this
515 It is used to optimize the build for a specific CPU / core that the
517 Generally, this does not change the ABI, though it can be a fine line
518 between optimization for specific cases.
522 in the top level Makefile for cross building.
523 Unused outside of that scope.
524 It is not passed down to the rest of the build.
525 Makefiles outside of the top level should not use it at all (though
526 some have their own private copy for hysterical raisons).
530 by the top level Makefile for cross building.
533 it is unused outside of that scope.
541 manual page appeared in