.\" Copyright (c) 2016-2017 The FreeBSD Foundation. All rights reserved.
.\"
.\" This documentation was created by Ed Maste under sponsorship of
.\" The FreeBSD Foundation.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\"    notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" $FreeBSD$
.\"
.Dd August 5, 2018
.Dt ARCH 7
.Os
.Sh NAME
.Nm arch
.Nd Architecture-specific details
.Sh DESCRIPTION
Differences between CPU architectures and platforms supported by
.Fx .
.Ss Introduction
This document is a quick reference of key ABI details of
.Fx
architecture ports.
For full details consult the processor-specific ABI supplement
documentation.
.Pp
If not explicitly mentioned, sizes are in bytes.
The architecture details in this document apply to
.Fx 10.0
and later, unless otherwise noted.
.Pp
.Fx
uses a flat address space.
Variables of types
.Vt unsigned long ,
.Vt uintptr_t ,
and
.Vt size_t
and pointers all have the same representation.
.Pp
In order to maximize compatibility with future pointer integrity mechanisms,
manipulations of pointers as integers should be performed via
.Vt uintptr_t
or
.Vt intptr_t
and no other types.
In particular,
.Vt long
and
.Vt ptrdiff_t
should be avoided.
.Pp
On some architectures, e.g.,
.Dv sparc64 ,
.Dv powerpc
and AIM variants of
.Dv powerpc64 ,
the kernel uses a separate address space.
On other architectures, kernel and a user mode process share a
single address space.
The kernel is located at the highest addresses.
.Pp
On each architecture, the main user mode thread's stack starts near
the highest user address and grows down.
.Pp
.Fx
architecture support varies by release.
This table shows the first
.Fx
release to support each architecture, and, for discontinued
architectures, the final release.
.Pp
.Bl -column -offset indent "Sy Architecture" "Sy Initial Release" "Sy Final Release"
.It Sy Architecture Ta Sy Initial Release Ta Sy Final Release
.It aarch64     Ta 11.0
.It alpha       Ta 3.2   Ta 6.4
.It amd64       Ta 5.1
.It arm         Ta 6.0
.It armeb       Ta 8.0   Ta 11.x
.It armv6       Ta 10.0
.It armv7       Ta 12.0
.It ia64        Ta 5.0   Ta 10.4
.It i386        Ta 1.0
.It mips        Ta 8.0
.It mipsel      Ta 9.0
.It mipselhf    Ta 12.0
.It mipshf      Ta 12.0
.It mipsn32     Ta 9.0
.It mips64      Ta 9.0
.It mips64el    Ta 9.0
.It mips64elhf  Ta 12.0
.It mips64hf    Ta 12.0
.It pc98        Ta 2.2   Ta 11.x
.It powerpc     Ta 6.0
.It powerpcspe  Ta 12.0
.It powerpc64   Ta 6.0
.It riscv64     Ta 12.0
.It riscv64sf   Ta 12.0
.It sparc64     Ta 5.0
.El
.Ss Type sizes
All
.Fx
architectures use some variant of the ELF (see
.Xr elf 5 )
.Sy Application Binary Interface
(ABI) for the machine processor.
All supported ABIs can be divided into two groups:
.Bl -tag -width "Dv ILP32"
.It Dv ILP32
.Vt int ,
.Vt long ,
.Vt void *
types machine representations all have 4-byte size.
.It Dv LP64
.Vt int
type machine representation uses 4 bytes,
while
.Vt long
and
.Vt void *
are 8 bytes.
.El
Compilers define the
.Dv _LP64
symbol when compiling for an
.Dv LP64
ABI.
.Pp
Some machines support more that one
.Fx
ABI.
Typically these are 64-bit machines, where the
.Dq native
.Dv LP64
execution environment is accompanied by the
.Dq legacy
.Dv ILP32
environment, which was historical 32-bit predecessor for 64-bit evolution.
Examples are:
.Bl -column -offset indent "Dv powerpc64" "Sy ILP32 counterpart"
.It Sy LP64        Ta Sy ILP32 counterpart
.It Dv amd64       Ta Dv i386
.It Dv powerpc64   Ta Dv powerpc
.It Dv mips64*     Ta Dv mips*
.El
.Dv aarch64
currently does not support execution of
.Dv armv6
or
.Dv armv7
binaries, even if the CPU implements
.Dv AArch32
execution state.
.Pp
On all supported architectures:
.Bl -column -offset -indent "long long" "Size"
.It Sy Type Ta Sy Size
.It short Ta 2
.It int Ta 4
.It long Ta sizeof(void*)
.It long long Ta 8
.It float Ta 4
.It double Ta 8
.El
Integers are represented in two's complement.
Alignment of integer and pointer types is natural, that is,
the address of the variable must be congruent to zero modulo the type size.
Most ILP32 ABIs, except
.Dv arm ,
require only 4-byte alignment for 64-bit integers.
.Pp
Machine-dependent type sizes:
.Bl -column -offset indent "Sy Architecture" "Sy void *" "Sy long double" "Sy time_t"
.It Sy Architecture Ta Sy void * Ta Sy long double Ta Sy time_t
.It aarch64     Ta 8 Ta 16 Ta 8
.It amd64       Ta 8 Ta 16 Ta 8
.It arm         Ta 4 Ta  8 Ta 8
.It armv6       Ta 4 Ta  8 Ta 8
.It i386        Ta 4 Ta 12 Ta 4
.It mips        Ta 4 Ta  8 Ta 8
.It mipsel      Ta 4 Ta  8 Ta 8
.It mipselhf    Ta 4 Ta  8 Ta 8
.It mipshf      Ta 4 Ta  8 Ta 8
.It mipsn32     Ta 4 Ta  8 Ta 8
.It mips64      Ta 8 Ta  8 Ta 8
.It mips64el    Ta 8 Ta  8 Ta 8
.It mips64elhf  Ta 8 Ta  8 Ta 8
.It mips64hf    Ta 8 Ta  8 Ta 8
.It powerpc     Ta 4 Ta  8 Ta 8
.It powerpcspe  Ta 4 Ta  8 Ta 8
.It powerpc64   Ta 8 Ta  8 Ta 8
.It riscv64     Ta 8 Ta 16 Ta 8
.It riscv64sf   Ta 8 Ta 16 Ta 8
.It sparc64     Ta 8 Ta 16 Ta 8
.El
.Pp
.Sy time_t
is 8 bytes on all supported architectures except i386.
.Ss Endianness and Char Signedness
.Bl -column -offset indent "Sy Architecture" "Sy Endianness" "Sy char Signedness"
.It Sy Architecture Ta Sy Endianness Ta Sy char Signedness
.It aarch64     Ta little Ta unsigned
.It amd64       Ta little Ta   signed
.It arm         Ta little Ta unsigned
.It armv6       Ta little Ta unsigned
.It armv7       Ta little Ta unsigned
.It i386        Ta little Ta   signed
.It mips        Ta big    Ta   signed
.It mipsel      Ta little Ta   signed
.It mipselhf    Ta little Ta   signed
.It mipshf      Ta big    Ta   signed
.It mipsn32     Ta big    Ta   signed
.It mips64      Ta big    Ta   signed
.It mips64el    Ta little Ta   signed
.It mips64elhf  Ta little Ta   signed
.It mips64hf    Ta big    Ta   signed
.It powerpc     Ta big    Ta unsigned
.It powerpcspe  Ta big    Ta unsigned
.It powerpc64   Ta big    Ta unsigned
.It riscv64     Ta little Ta   signed
.It riscv64sf   Ta little Ta   signed
.It sparc64     Ta big    Ta   signed
.El
.Ss Page Size
.Bl -column -offset indent "Sy Architecture" "Sy Page Sizes"
.It Sy Architecture Ta Sy Page Sizes
.It aarch64     Ta 4K, 2M, 1G
.It amd64       Ta 4K, 2M, 1G
.It arm         Ta 4K
.It armv6       Ta 4K, 1M
.It armv7       Ta 4K, 1M
.It i386        Ta 4K, 2M (PAE), 4M
.It mips        Ta 4K
.It mipsel      Ta 4K
.It mipselhf    Ta 4K
.It mipshf      Ta 4K
.It mipsn32     Ta 4K
.It mips64      Ta 4K
.It mips64el    Ta 4K
.It mips64elhf  Ta 4K
.It mips64hf    Ta 4K
.It powerpc     Ta 4K
.It powerpcspe  Ta 4K
.It powerpc64   Ta 4K
.It riscv64     Ta 4K
.It riscv64sf   Ta 4K
.It sparc64     Ta 8K
.El
.Ss Floating Point
.Bl -column -offset indent "Sy Architecture" "Sy float, double" "Sy long double"
.It Sy Architecture Ta Sy float, double Ta Sy long double
.It aarch64     Ta hard Ta soft, quad precision
.It amd64       Ta hard Ta hard, 80 bit
.It arm         Ta soft Ta soft, double precision
.It armv6       Ta hard(1) Ta hard, double precision
.It armv7       Ta hard(1) Ta hard, double precision
.It i386        Ta hard Ta hard, 80 bit
.It mips        Ta soft Ta identical to double
.It mipsel      Ta soft Ta identical to double
.It mipselhf    Ta hard Ta identical to double
.It mipshf      Ta hard Ta identical to double
.It mipsn32     Ta soft Ta identical to double
.It mips64      Ta soft Ta identical to double
.It mips64el    Ta soft Ta identical to double
.It mips64elhf  Ta hard Ta identical to double
.It mips64hf    Ta hard Ta identical to double
.It powerpc     Ta hard Ta hard, double precision
.It powerpcspe  Ta hard Ta hard, double precision
.It powerpc64   Ta hard Ta hard, double precision
.It riscv64     Ta hard Ta hard, double precision
.It riscv64sf   Ta soft Ta soft, double precision
.It sparc64     Ta hard Ta hard, quad precision
.El
.Pp
(1) Prior to
.Fx 11.0 ,
armv6 used the softfp ABI even though it supported only processors
with a floating point unit.
.Ss Predefined Macros
The compiler provides a number of predefined macros.
Some of these provide architecture-specific details and are explained below.
Other macros, including those required by the language standard, are not
included here.
.Pp
The full set of predefined macros can be obtained with this command:
.Bd -literal -offset indent
cc -x c -dM -E /dev/null
.Ed
.Pp
Common type size and endianness macros:
.Bl -column -offset indent "BYTE_ORDER" "Sy Meaning"
.It Sy Macro Ta Sy Meaning
.It Dv __LP64__ Ta 64-bit (8-byte) long and pointer, 32-bit (4-byte) int
.It Dv __ILP32__ Ta 32-bit (4-byte) int, long and pointer
.It Dv BYTE_ORDER Ta Either Dv BIG_ENDIAN or Dv LITTLE_ENDIAN .
.Dv PDP11_ENDIAN
is not used on
.Fx .
.El
.Pp
Architecture-specific macros:
.Bl -column -offset indent "Sy Architecture" "Sy Predefined macros"
.It Sy Architecture Ta Sy Predefined macros
.It aarch64     Ta Dv __aarch64__
.It amd64       Ta Dv __amd64__, Dv __x86_64__
.It arm         Ta Dv __arm__
.It armv6       Ta Dv __arm__, Dv __ARM_ARCH >= 6
.It armv7       Ta Dv __arm__, Dv __ARM_ARCH >= 7
.It i386        Ta Dv __i386__
.It mips        Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_o32
.It mipsel      Ta Dv __mips__, Dv __mips_o32
.It mipselhf    Ta Dv __mips__, Dv __mips_o32
.It mipshf      Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_o32
.It mipsn32     Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n32
.It mips64      Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n64
.It mips64el    Ta Dv __mips__, Dv __mips_n64
.It mips64elhf  Ta Dv __mips__, Dv __mips_n64
.It mips64hf    Ta Dv __mips__, Dv __MIPSEB__, Dv __mips_n64
.It powerpc     Ta Dv __powerpc__
.It powerpcspe  Ta Dv __powerpc__, Dv __SPE__
.It powerpc64   Ta Dv __powerpc__, Dv __powerpc64__
.It riscv64     Ta Dv __riscv, Dv __riscv_xlen == 64
.It riscv64sf   Ta Dv __riscv, Dv __riscv_xlen == 64
.It sparc64     Ta Dv __sparc64__
.El
.Pp
Compilers may define additional variants of architecture-specific macros.
The macros above are preferred for use in
.Fx .
.Ss Important Xr make 1 variables
Most of the externally settable variables are defined in the
.Xr build 7
man page.
These variables are not otherwise documented and are used extensively
in the build system.
.Bl -column -offset indent "Sy Variable" "Sy Meaning and usage"
.It Dv MACHINE	Represent the hardware platform.
This is the same as the native platform's
.Xr uname 1
.Fl m
output.
It defines both the userland / kernel interface, as well as the
bootloader / kernel interface.
It should only be used in these contexts.
Each CPU architecture may have multiple hardware platforms it supports
where
.Dv MACHINE
differs among them.
It is used to collect together all the files from
.Xr config 8
to build the kernel.
It is often the same as
.Dv MACHINE_ARCH
just as one CPU architecture can be implemented by many different
hardware platforms, one hardware platform may support multiple CPU
architecture family members, though with different binaries.
For example,
.Dv MACHINE
of i386 supported the IBM-AT hardware platform while the
.Dv MACHINE
of pc98 supported the Japanese company NEC's PC-9801 and PC-9821
hardware platforms.
Both of these hardware platforms supported only the
.Dv MACHINE_ARCH
of i386 where they shared a common ABI, except for certain kernel /
userland interfaces relating to underlying hardware platform
differences in bus architecture, device enumeration and boot interface.
Generally,
.Dv MACHINE
should only be used in src/sys and src/stand or in system imagers or
installers.
.It Dv MACHINE_ARCH	Represents the CPU processor architecture.
This is the same as the native platforms
.Xr uname 1
.Fl p
output.
It defines the CPU instruction family supported.
It may also encode a variation in the byte ordering of multi-byte
integers (endian).
It may also encode a variation in the size of the integer or pointer.
It may also encode a ISA revision.
It may also encode hard versus soft floating point ABI and usage.
It may also encode a variant ABI when the other factors do not
uniquely define the ABI (e.g., MIPS' n32 ABI).
It, along with
.Dv MACHINE ,
defines the ABI used by the system.
For example, the MIPS CPU processor family supports 9 different
combinations encoding pointer size, endian and hard versus soft float (for
8 combinations) as well as N32 (which only ever had one variation of
all these).
Generally, the plain CPU name specifies the most common (or at least
first) variant of the CPU.
This is why mips and mips64 imply 'big endian' while 'arm' and 'armv7'
imply little endian.
If we ever were to support the so-called x32 ABI (using 32-bit
pointers on the amd64 architecture), it would most likely be encoded
as amd64-x32.
It is unfortunate that amd64 specifies the 64-bit evolution of the x86
platform (it matches the 'first rule') as everybody else uses x86_64.
There is no standard name for the processor: each OS selects its own
conventions.
.It Dv MACHINE_CPUARCH	Represents the source location for a given
.Dv MACHINE_ARCH .
For example,
.Dv MACHINE_CPUARCH
is defined to be mips for all the flavors of mips that we support
since we support them all with a shared set of sources.
While amd64 and i386 are closely related, MACHINE_CPUARCH is not x86
for them.
The FreeBSD source base supports amd64 and i386 with two
distinct source bases living in subdirectories named amd64 and i386
(though behind the scenes there's some sharing that fits into this
framework).
.It Dv CPUTYPE	Sets the flavor of
.Dv MACHINE_ARCH
to build.
It is used to optimize the build for a specific CPU / core that the
binaries run on.
Generally, this does not change the ABI, though it can be a fine line
between optimization for specific cases.
.It Dv TARGET	Used to set
.Dv MACHINE
in the top level Makefile for cross building.
Unused outside of that scope.
It is not passed down to the rest of the build.
Makefiles outside of the top level should not use it at all (though
some have their own private copy for hysterical raisons).
.It Dv TARGET_ARCH	Used to set
.Dv MACHINE_ARCH
by the top level Makefile for cross building.
Like
.Dv TARGET , it is unused outside of that scope.
.El
.Sh SEE ALSO
.Xr src.conf 5 ,
.Xr build 7
.Sh HISTORY
An
.Nm
manual page appeared in
.Fx 11.1 .