1.\" Copyright (c) 2016-2017 The FreeBSD Foundation. All rights reserved. 2.\" 3.\" This documentation was created by Ed Maste under sponsorship of 4.\" The FreeBSD Foundation. 5.\" 6.\" Redistribution and use in source and binary forms, with or without 7.\" modification, are permitted provided that the following conditions 8.\" are met: 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. 14.\" 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 25.\" SUCH DAMAGE. 26.\" 27.\" $FreeBSD$ 28.\" 29.Dd August 5, 2018 30.Dt ARCH 7 31.Os 32.Sh NAME 33.Nm arch 34.Nd Architecture-specific details 35.Sh DESCRIPTION 36Differences between CPU architectures and platforms supported by 37.Fx . 38.Ss Introduction 39This document is a quick reference of key ABI details of 40.Fx 41architecture ports. 42For full details consult the processor-specific ABI supplement 43documentation. 44.Pp 45If not explicitly mentioned, sizes are in bytes. 46The architecture details in this document apply to 47.Fx 10.0 48and later, unless otherwise noted. 49.Pp 50.Fx 51uses a flat address space. 52Variables of types 53.Vt unsigned long , 54.Vt uintptr_t , 55and 56.Vt size_t 57and pointers all have the same representation. 58.Pp 59In order to maximize compatibility with future pointer integrity mechanisms, 60manipulations of pointers as integers should be performed via 61.Vt uintptr_t 62or 63.Vt intptr_t 64and no other types. 65In particular, 66.Vt long 67and 68.Vt ptrdiff_t 69should be avoided. 70.Pp 71On some architectures, e.g., 72.Dv sparc64 , 73.Dv powerpc 74and AIM variants of 75.Dv powerpc64 , 76the kernel uses a separate address space. 77On other architectures, kernel and a user mode process share a 78single address space. 79The kernel is located at the highest addresses. 80.Pp 81On each architecture, the main user mode thread's stack starts near 82the highest user address and grows down. 83.Pp 84.Fx 85architecture support varies by release. 86This table shows the first 87.Fx 88release to support each architecture, and, for discontinued 89architectures, the final release. 90.Pp 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 93.It aarch64 Ta 11.0 94.It alpha Ta 3.2 Ta 6.4 95.It amd64 Ta 5.1 96.It arm Ta 6.0 97.It armeb Ta 8.0 Ta 11.x 98.It armv6 Ta 10.0 99.It armv7 Ta 12.0 100.It ia64 Ta 5.0 Ta 10.4 101.It i386 Ta 1.0 102.It mips Ta 8.0 103.It mipsel Ta 9.0 104.It mipselhf Ta 12.0 105.It mipshf Ta 12.0 106.It mipsn32 Ta 9.0 107.It mips64 Ta 9.0 108.It mips64el Ta 9.0 109.It mips64elhf Ta 12.0 110.It mips64hf Ta 12.0 111.It pc98 Ta 2.2 Ta 11.x 112.It powerpc Ta 6.0 113.It powerpcspe Ta 12.0 114.It powerpc64 Ta 6.0 115.It riscv64 Ta 12.0 116.It riscv64sf Ta 12.0 117.It sparc64 Ta 5.0 118.El 119.Ss Type sizes 120All 121.Fx 122architectures use some variant of the ELF (see 123.Xr elf 5 ) 124.Sy Application Binary Interface 125(ABI) for the machine processor. 126All supported ABIs can be divided into two groups: 127.Bl -tag -width "Dv ILP32" 128.It Dv ILP32 129.Vt int , 130.Vt long , 131.Vt void * 132types machine representations all have 4-byte size. 133.It Dv LP64 134.Vt int 135type machine representation uses 4 bytes, 136while 137.Vt long 138and 139.Vt void * 140are 8 bytes. 141.El 142Compilers define the 143.Dv _LP64 144symbol when compiling for an 145.Dv LP64 146ABI. 147.Pp 148Some machines support more that one 149.Fx 150ABI. 151Typically these are 64-bit machines, where the 152.Dq native 153.Dv LP64 154execution environment is accompanied by the 155.Dq legacy 156.Dv ILP32 157environment, which was historical 32-bit predecessor for 64-bit evolution. 158Examples are: 159.Bl -column -offset indent "Dv powerpc64" "Sy ILP32 counterpart" 160.It Sy LP64 Ta Sy ILP32 counterpart 161.It Dv amd64 Ta Dv i386 162.It Dv powerpc64 Ta Dv powerpc 163.It Dv mips64* Ta Dv mips* 164.El 165.Dv aarch64 166currently does not support execution of 167.Dv armv6 168or 169.Dv armv7 170binaries, even if the CPU implements 171.Dv AArch32 172execution state. 173.Pp 174On all supported architectures: 175.Bl -column -offset -indent "long long" "Size" 176.It Sy Type Ta Sy Size 177.It short Ta 2 178.It int Ta 4 179.It long Ta sizeof(void*) 180.It long long Ta 8 181.It float Ta 4 182.It double Ta 8 183.El 184Integers are represented in two's complement. 185Alignment of integer and pointer types is natural, that is, 186the address of the variable must be congruent to zero modulo the type size. 187Most ILP32 ABIs, except 188.Dv arm , 189require only 4-byte alignment for 64-bit integers. 190.Pp 191Machine-dependent type sizes: 192.Bl -column -offset indent "Sy Architecture" "Sy void *" "Sy long double" "Sy time_t" 193.It Sy Architecture Ta Sy void * Ta Sy long double Ta Sy time_t 194.It aarch64 Ta 8 Ta 16 Ta 8 195.It amd64 Ta 8 Ta 16 Ta 8 196.It arm Ta 4 Ta 8 Ta 8 197.It armv6 Ta 4 Ta 8 Ta 8 198.It i386 Ta 4 Ta 12 Ta 4 199.It mips Ta 4 Ta 8 Ta 8 200.It mipsel Ta 4 Ta 8 Ta 8 201.It mipselhf Ta 4 Ta 8 Ta 8 202.It mipshf Ta 4 Ta 8 Ta 8 203.It mipsn32 Ta 4 Ta 8 Ta 8 204.It mips64 Ta 8 Ta 8 Ta 8 205.It mips64el Ta 8 Ta 8 Ta 8 206.It mips64elhf Ta 8 Ta 8 Ta 8 207.It mips64hf Ta 8 Ta 8 Ta 8 208.It powerpc Ta 4 Ta 8 Ta 8 209.It powerpcspe Ta 4 Ta 8 Ta 8 210.It powerpc64 Ta 8 Ta 8 Ta 8 211.It riscv64 Ta 8 Ta 16 Ta 8 212.It riscv64sf Ta 8 Ta 16 Ta 8 213.It sparc64 Ta 8 Ta 16 Ta 8 214.El 215.Pp 216.Sy time_t 217is 8 bytes on all supported architectures except i386. 218.Ss Endianness and Char Signedness 219.Bl -column -offset indent "Sy Architecture" "Sy Endianness" "Sy char Signedness" 220.It Sy Architecture Ta Sy Endianness Ta Sy char Signedness 221.It aarch64 Ta little Ta unsigned 222.It amd64 Ta little Ta signed 223.It arm Ta little Ta unsigned 224.It armv6 Ta little Ta unsigned 225.It armv7 Ta little Ta unsigned 226.It i386 Ta little Ta signed 227.It mips Ta big Ta signed 228.It mipsel Ta little Ta signed 229.It mipselhf Ta little Ta signed 230.It mipshf Ta big Ta signed 231.It mipsn32 Ta big Ta signed 232.It mips64 Ta big Ta signed 233.It mips64el Ta little Ta signed 234.It mips64elhf Ta little Ta signed 235.It mips64hf Ta big Ta signed 236.It powerpc Ta big Ta unsigned 237.It powerpcspe Ta big Ta unsigned 238.It powerpc64 Ta big Ta unsigned 239.It riscv64 Ta little Ta signed 240.It riscv64sf Ta little Ta signed 241.It sparc64 Ta big Ta signed 242.El 243.Ss Page Size 244.Bl -column -offset indent "Sy Architecture" "Sy Page Sizes" 245.It Sy Architecture Ta Sy Page Sizes 246.It aarch64 Ta 4K, 2M, 1G 247.It amd64 Ta 4K, 2M, 1G 248.It arm Ta 4K 249.It armv6 Ta 4K, 1M 250.It armv7 Ta 4K, 1M 251.It i386 Ta 4K, 2M (PAE), 4M 252.It mips Ta 4K 253.It mipsel Ta 4K 254.It mipselhf Ta 4K 255.It mipshf Ta 4K 256.It mipsn32 Ta 4K 257.It mips64 Ta 4K 258.It mips64el Ta 4K 259.It mips64elhf Ta 4K 260.It mips64hf Ta 4K 261.It powerpc Ta 4K 262.It powerpcspe Ta 4K 263.It powerpc64 Ta 4K 264.It riscv64 Ta 4K 265.It riscv64sf Ta 4K 266.It sparc64 Ta 8K 267.El 268.Ss Floating Point 269.Bl -column -offset indent "Sy Architecture" "Sy float, double" "Sy long double" 270.It Sy Architecture Ta Sy float, double Ta Sy long double 271.It aarch64 Ta hard Ta soft, quad precision 272.It amd64 Ta hard Ta hard, 80 bit 273.It arm Ta soft Ta soft, double precision 274.It armv6 Ta hard(1) Ta hard, double precision 275.It armv7 Ta hard(1) Ta hard, double precision 276.It i386 Ta hard Ta hard, 80 bit 277.It mips Ta soft Ta identical to double 278.It mipsel Ta soft Ta identical to double 279.It mipselhf Ta hard Ta identical to double 280.It mipshf Ta hard Ta identical to double 281.It mipsn32 Ta soft Ta identical to double 282.It mips64 Ta soft Ta identical to double 283.It mips64el Ta soft Ta identical to double 284.It mips64elhf Ta hard Ta identical to double 285.It mips64hf Ta hard Ta identical to double 286.It powerpc Ta hard Ta hard, double precision 287.It powerpcspe Ta hard Ta hard, double precision 288.It powerpc64 Ta hard Ta hard, double precision 289.It riscv64 Ta hard Ta hard, double precision 290.It riscv64sf Ta soft Ta soft, double precision 291.It sparc64 Ta hard Ta hard, quad precision 292.El 293.Pp 294(1) Prior to 295.Fx 11.0 , 296armv6 used the softfp ABI even though it supported only processors 297with a floating point unit. 298.Ss Predefined Macros 299The compiler provides a number of predefined macros. 300Some of these provide architecture-specific details and are explained below. 301Other macros, including those required by the language standard, are not 302included here. 303.Pp 304The full set of predefined macros can be obtained with this command: 305.Bd -literal -offset indent 306cc -x c -dM -E /dev/null 307.Ed 308.Pp 309Common type size and endianness macros: 310.Bl -column -offset indent "BYTE_ORDER" "Sy Meaning" 311.It Sy Macro Ta Sy Meaning 312.It Dv __LP64__ Ta 64-bit (8-byte) long and pointer, 32-bit (4-byte) int 313.It Dv __ILP32__ Ta 32-bit (4-byte) int, long and pointer 314.It Dv BYTE_ORDER Ta Either Dv BIG_ENDIAN or Dv LITTLE_ENDIAN . 315.Dv PDP11_ENDIAN 316is not used on 317.Fx . 318.El 319.Pp 320Architecture-specific macros: 321.Bl -column -offset indent "Sy Architecture" "Sy Predefined macros" 322.It Sy Architecture Ta Sy Predefined macros 323.It aarch64 Ta Dv __aarch64__ 324.It amd64 Ta Dv __amd64__, Dv __x86_64__ 325.It arm Ta Dv __arm__ 326.It armv6 Ta Dv __arm__, Dv __ARM_ARCH >= 6 327.It armv7 Ta Dv __arm__, Dv __ARM_ARCH >= 7 328.It i386 Ta Dv __i386__ 329.It mips Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_o32 330.It mipsel Ta Dv __mips__, Dv __mips_o32 331.It mipselhf Ta Dv __mips__, Dv __mips_o32 332.It mipshf Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_o32 333.It mipsn32 Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n32 334.It mips64 Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n64 335.It mips64el Ta Dv __mips__, Dv __mips_n64 336.It mips64elhf Ta Dv __mips__, Dv __mips_n64 337.It mips64hf Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n64 338.It powerpc Ta Dv __powerpc__ 339.It powerpcspe Ta Dv __powerpc__, Dv __SPE__ 340.It powerpc64 Ta Dv __powerpc__, Dv __powerpc64__ 341.It riscv64 Ta Dv __riscv, Dv __riscv_xlen == 64 342.It riscv64sf Ta Dv __riscv, Dv __riscv_xlen == 64 343.It sparc64 Ta Dv __sparc64__ 344.El 345.Pp 346Compilers may define additional variants of architecture-specific macros. 347The macros above are preferred for use in 348.Fx . 349.Ss Important Xr make 1 variables 350Most of the externally settable variables are defined in the 351.Xr build 7 352man page. 353These variables are not otherwise documented and are used extensively 354in the build system. 355.Bl -column -offset indent "Sy Variable" "Sy Meaning and usage" 356.It Dv MACHINE Represent the hardware platform. 357This is the same as the native platform's 358.Xr uname 1 359.Fl m 360output. 361It defines both the userland / kernel interface, as well as the 362bootloader / kernel interface. 363It should only be used in these contexts. 364Each CPU architecture may have multiple hardware platforms it supports 365where 366.Dv MACHINE 367differs among them. 368It is used to collect together all the files from 369.Xr config 8 370to build the kernel. 371It is often the same as 372.Dv MACHINE_ARCH 373just as one CPU architecture can be implemented by many different 374hardware platforms, one hardware platform may support multiple CPU 375architecture family members, though with different binaries. 376For example, 377.Dv MACHINE 378of i386 supported the IBM-AT hardware platform while the 379.Dv MACHINE 380of pc98 supported the Japanese company NEC's PC-9801 and PC-9821 381hardware platforms. 382Both of these hardware platforms supported only the 383.Dv MACHINE_ARCH 384of i386 where they shared a common ABI, except for certain kernel / 385userland interfaces relating to underlying hardware platform 386differences in bus architecture, device enumeration and boot interface. 387Generally, 388.Dv MACHINE 389should only be used in src/sys and src/stand or in system imagers or 390installers. 391.It Dv MACHINE_ARCH Represents the CPU processor architecture. 392This is the same as the native platforms 393.Xr uname 1 394.Fl p 395output. 396It defines the CPU instruction family supported. 397It may also encode a variation in the byte ordering of multi-byte 398integers (endian). 399It may also encode a variation in the size of the integer or pointer. 400It may also encode a ISA revision. 401It may also encode hard versus soft floating point ABI and usage. 402It may also encode a variant ABI when the other factors do not 403uniquely define the ABI (e.g., MIPS' n32 ABI). 404It, along with 405.Dv MACHINE , 406defines the ABI used by the system. 407For example, the MIPS CPU processor family supports 9 different 408combinations encoding pointer size, endian and hard versus soft float (for 4098 combinations) as well as N32 (which only ever had one variation of 410all these). 411Generally, the plain CPU name specifies the most common (or at least 412first) variant of the CPU. 413This is why mips and mips64 imply 'big endian' while 'arm' and 'armv7' 414imply little endian. 415If we ever were to support the so-called x32 ABI (using 32-bit 416pointers on the amd64 architecture), it would most likely be encoded 417as amd64-x32. 418It is unfortunate that amd64 specifies the 64-bit evolution of the x86 419platform (it matches the 'first rule') as everybody else uses x86_64. 420There is no standard name for the processor: each OS selects its own 421conventions. 422.It Dv MACHINE_CPUARCH Represents the source location for a given 423.Dv MACHINE_ARCH . 424For example, 425.Dv MACHINE_CPUARCH 426is defined to be mips for all the flavors of mips that we support 427since we support them all with a shared set of sources. 428While amd64 and i386 are closely related, MACHINE_CPUARCH is not x86 429for them. 430The FreeBSD source base supports amd64 and i386 with two 431distinct source bases living in subdirectories named amd64 and i386 432(though behind the scenes there's some sharing that fits into this 433framework). 434.It Dv CPUTYPE Sets the flavor of 435.Dv MACHINE_ARCH 436to build. 437It is used to optimize the build for a specific CPU / core that the 438binaries run on. 439Generally, this does not change the ABI, though it can be a fine line 440between optimization for specific cases. 441.It Dv TARGET Used to set 442.Dv MACHINE 443in the top level Makefile for cross building. 444Unused outside of that scope. 445It is not passed down to the rest of the build. 446Makefiles outside of the top level should not use it at all (though 447some have their own private copy for hysterical raisons). 448.It Dv TARGET_ARCH Used to set 449.Dv MACHINE_ARCH 450by the top level Makefile for cross building. 451Like 452.Dv TARGET , it is unused outside of that scope. 453.El 454.Sh SEE ALSO 455.Xr src.conf 5 , 456.Xr build 7 457.Sh HISTORY 458An 459.Nm 460manual page appeared in 461.Fx 11.1 . 462