📄 machine-constraints.html
字号:
<html lang="en">
<head>
<title>Using the GNU Compiler Collection (GCC)</title>
<meta http-equiv="Content-Type" content="text/html">
<meta name="description" content="Using the GNU Compiler Collection (GCC)">
<meta name="generator" content="makeinfo 4.3">
<link href="http://www.gnu.org/software/texinfo/" rel="generator-home">
<!--
Copyright © 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
1999, 2000, 2001, 2002, 2003 Free Software Foundation, Inc.
<p>Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.2 or
any later version published by the Free Software Foundation; with the
Invariant Sections being "GNU General Public License" and "Funding
Free Software", the Front-Cover texts being (a) (see below), and with
the Back-Cover Texts being (b) (see below). A copy of the license is
included in the section entitled "GNU Free Documentation License".
<p>(a) The FSF's Front-Cover Text is:
<p>A GNU Manual
<p>(b) The FSF's Back-Cover Text is:
<p>You have freedom to copy and modify this GNU Manual, like GNU
software. Copies published by the Free Software Foundation raise
funds for GNU development.-->
</head>
<body>
<div class="node">
<p>
Node:<a name="Machine%20Constraints">Machine Constraints</a>,
Previous:<a rel="previous" accesskey="p" href="Modifiers.html#Modifiers">Modifiers</a>,
Up:<a rel="up" accesskey="u" href="Constraints.html#Constraints">Constraints</a>
<hr><br>
</div>
<h4 class="subsection">Constraints for Particular Machines</h4>
<p>Whenever possible, you should use the general-purpose constraint letters
in <code>asm</code> arguments, since they will convey meaning more readily to
people reading your code. Failing that, use the constraint letters
that usually have very similar meanings across architectures. The most
commonly used constraints are <code>m</code> and <code>r</code> (for memory and
general-purpose registers respectively; see <a href="Simple-Constraints.html#Simple%20Constraints">Simple Constraints</a>), and
<code>I</code>, usually the letter indicating the most common
immediate-constant format.
<p>For each machine architecture, the
<code>config/</code><var>machine</var><code>/</code><var>machine</var><code>.h</code> file defines additional
constraints. These constraints are used by the compiler itself for
instruction generation, as well as for <code>asm</code> statements; therefore,
some of the constraints are not particularly interesting for <code>asm</code>.
The constraints are defined through these macros:
<dl>
<dt><code>REG_CLASS_FROM_LETTER</code>
<dd>Register class constraints (usually lower case).
<br><dt><code>CONST_OK_FOR_LETTER_P</code>
<dd>Immediate constant constraints, for non-floating point constants of
word size or smaller precision (usually upper case).
<br><dt><code>CONST_DOUBLE_OK_FOR_LETTER_P</code>
<dd>Immediate constant constraints, for all floating point constants and for
constants of greater than word size precision (usually upper case).
<br><dt><code>EXTRA_CONSTRAINT</code>
<dd>Special cases of registers or memory. This macro is not required, and
is only defined for some machines.
</dl>
<p>Inspecting these macro definitions in the compiler source for your
machine is the best way to be certain you have the right constraints.
However, here is a summary of the machine-dependent constraints
available on some particular machines.
<dl>
<dt><em>ARM family--</em><code>arm.h</code><em></em>
<dd>
<dl>
<dt><code>f</code>
<dd>Floating-point register
<br><dt><code>F</code>
<dd>One of the floating-point constants 0.0, 0.5, 1.0, 2.0, 3.0, 4.0, 5.0
or 10.0
<br><dt><code>G</code>
<dd>Floating-point constant that would satisfy the constraint <code>F</code> if it
were negated
<br><dt><code>I</code>
<dd>Integer that is valid as an immediate operand in a data processing
instruction. That is, an integer in the range 0 to 255 rotated by a
multiple of 2
<br><dt><code>J</code>
<dd>Integer in the range -4095 to 4095
<br><dt><code>K</code>
<dd>Integer that satisfies constraint <code>I</code> when inverted (ones complement)
<br><dt><code>L</code>
<dd>Integer that satisfies constraint <code>I</code> when negated (twos complement)
<br><dt><code>M</code>
<dd>Integer in the range 0 to 32
<br><dt><code>Q</code>
<dd>A memory reference where the exact address is in a single register
(`<code>m</code>' is preferable for <code>asm</code> statements)
<br><dt><code>R</code>
<dd>An item in the constant pool
<br><dt><code>S</code>
<dd>A symbol in the text segment of the current file
</dl>
<br><dt><em>AVR family--</em><code>avr.h</code><em></em>
<dd>
<dl>
<dt><code>l</code>
<dd>Registers from r0 to r15
<br><dt><code>a</code>
<dd>Registers from r16 to r23
<br><dt><code>d</code>
<dd>Registers from r16 to r31
<br><dt><code>w</code>
<dd>Registers from r24 to r31. These registers can be used in <code>adiw</code> command
<br><dt><code>e</code>
<dd>Pointer register (r26-r31)
<br><dt><code>b</code>
<dd>Base pointer register (r28-r31)
<br><dt><code>q</code>
<dd>Stack pointer register (SPH:SPL)
<br><dt><code>t</code>
<dd>Temporary register r0
<br><dt><code>x</code>
<dd>Register pair X (r27:r26)
<br><dt><code>y</code>
<dd>Register pair Y (r29:r28)
<br><dt><code>z</code>
<dd>Register pair Z (r31:r30)
<br><dt><code>I</code>
<dd>Constant greater than -1, less than 64
<br><dt><code>J</code>
<dd>Constant greater than -64, less than 1
<br><dt><code>K</code>
<dd>Constant integer 2
<br><dt><code>L</code>
<dd>Constant integer 0
<br><dt><code>M</code>
<dd>Constant that fits in 8 bits
<br><dt><code>N</code>
<dd>Constant integer -1
<br><dt><code>O</code>
<dd>Constant integer 8, 16, or 24
<br><dt><code>P</code>
<dd>Constant integer 1
<br><dt><code>G</code>
<dd>A floating point constant 0.0
</dl>
<br><dt><em>IBM RS6000--</em><code>rs6000.h</code><em></em>
<dd>
<dl>
<dt><code>b</code>
<dd>Address base register
<br><dt><code>f</code>
<dd>Floating point register
<br><dt><code>h</code>
<dd><code>MQ</code>, <code>CTR</code>, or <code>LINK</code> register
<br><dt><code>q</code>
<dd><code>MQ</code> register
<br><dt><code>c</code>
<dd><code>CTR</code> register
<br><dt><code>l</code>
<dd><code>LINK</code> register
<br><dt><code>x</code>
<dd><code>CR</code> register (condition register) number 0
<br><dt><code>y</code>
<dd><code>CR</code> register (condition register)
<br><dt><code>z</code>
<dd><code>FPMEM</code> stack memory for FPR-GPR transfers
<br><dt><code>I</code>
<dd>Signed 16-bit constant
<br><dt><code>J</code>
<dd>Unsigned 16-bit constant shifted left 16 bits (use <code>L</code> instead for
<code>SImode</code> constants)
<br><dt><code>K</code>
<dd>Unsigned 16-bit constant
<br><dt><code>L</code>
<dd>Signed 16-bit constant shifted left 16 bits
<br><dt><code>M</code>
<dd>Constant larger than 31
<br><dt><code>N</code>
<dd>Exact power of 2
<br><dt><code>O</code>
<dd>Zero
<br><dt><code>P</code>
<dd>Constant whose negation is a signed 16-bit constant
<br><dt><code>G</code>
<dd>Floating point constant that can be loaded into a register with one
instruction per word
<br><dt><code>Q</code>
<dd>Memory operand that is an offset from a register (<code>m</code> is preferable
for <code>asm</code> statements)
<br><dt><code>R</code>
<dd>AIX TOC entry
<br><dt><code>S</code>
<dd>Constant suitable as a 64-bit mask operand
<br><dt><code>T</code>
<dd>Constant suitable as a 32-bit mask operand
<br><dt><code>U</code>
<dd>System V Release 4 small data area reference
</dl>
<br><dt><em>Intel 386--</em><code>i386.h</code><em></em>
<dd>
<dl>
<dt><code>q</code>
<dd><code>a</code>, <code>b</code>, <code>c</code>, or <code>d</code> register for the i386.
For x86-64 it is equivalent to <code>r</code> class. (for 8-bit instructions that
do not use upper halves)
<br><dt><code>Q</code>
<dd><code>a</code>, <code>b</code>, <code>c</code>, or <code>d</code> register. (for 8-bit instructions,
that do use upper halves)
<br><dt><code>R</code>
<dd>Legacy register--equivalent to <code>r</code> class in i386 mode.
(for non-8-bit registers used together with 8-bit upper halves in a single
instruction)
<br><dt><code>A</code>
<dd>Specifies the <code>a</code> or <code>d</code> registers. This is primarily useful
for 64-bit integer values (when in 32-bit mode) intended to be returned
with the <code>d</code> register holding the most significant bits and the
<code>a</code> register holding the least significant bits.
<br><dt><code>f</code>
<dd>Floating point register
<br><dt><code>t</code>
<dd>First (top of stack) floating point register
<br><dt><code>u</code>
<dd>Second floating point register
<br><dt><code>a</code>
<dd><code>a</code> register
<br><dt><code>b</code>
<dd><code>b</code> register
<br><dt><code>c</code>
<dd><code>c</code> register
<br><dt><code>C</code>
<dd>Specifies constant that can be easily constructed in SSE register without
loading it from memory.
<br><dt><code>d</code>
<dd><code>d</code> register
<br><dt><code>D</code>
<dd><code>di</code> register
<br><dt><code>S</code>
<dd><code>si</code> register
<br><dt><code>x</code>
<dd><code>xmm</code> SSE register
<br><dt><code>y</code>
<dd>MMX register
<br><dt><code>I</code>
<dd>Constant in range 0 to 31 (for 32-bit shifts)
<br><dt><code>J</code>
<dd>Constant in range 0 to 63 (for 64-bit shifts)
<br><dt><code>K</code>
<dd><code>0xff</code>
<br><dt><code>L</code>
<dd><code>0xffff</code>
<br><dt><code>M</code>
<dd>0, 1, 2, or 3 (shifts for <code>lea</code> instruction)
<br><dt><code>N</code>
<dd>Constant in range 0 to 255 (for <code>out</code> instruction)
<br><dt><code>Z</code>
<dd>Constant in range 0 to <code>0xffffffff</code> or symbolic reference known to fit specified range.
(for using immediates in zero extending 32-bit to 64-bit x86-64 instructions)
<br><dt><code>e</code>
<dd>Constant in range -2147483648 to 2147483647 or symbolic reference known to fit specified range.
(for using immediates in 64-bit x86-64 instructions)
<br><dt><code>G</code>
<dd>Standard 80387 floating point constant
</dl>
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -