1<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook V3.1//EN"[]> 2 3<book id="DoingIO"> 4 <bookinfo> 5 <title>Bus-Independent Device Accesses</title> 6 7 <authorgroup> 8 <author> 9 <firstname>Matthew</firstname> 10 <surname>Wilcox</surname> 11 <affiliation> 12 <address> 13 <email>matthew@wil.cx</email> 14 </address> 15 </affiliation> 16 </author> 17 </authorgroup> 18 19 <authorgroup> 20 <author> 21 <firstname>Alan</firstname> 22 <surname>Cox</surname> 23 <affiliation> 24 <address> 25 <email>alan@redhat.com</email> 26 </address> 27 </affiliation> 28 </author> 29 </authorgroup> 30 31 <copyright> 32 <year>2001</year> 33 <holder>Matthew Wilcox</holder> 34 </copyright> 35 36 <legalnotice> 37 <para> 38 This documentation is free software; you can redistribute 39 it and/or modify it under the terms of the GNU General Public 40 License as published by the Free Software Foundation; either 41 version 2 of the License, or (at your option) any later 42 version. 43 </para> 44 45 <para> 46 This program is distributed in the hope that it will be 47 useful, but WITHOUT ANY WARRANTY; without even the implied 48 warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. 49 See the GNU General Public License for more details. 50 </para> 51 52 <para> 53 You should have received a copy of the GNU General Public 54 License along with this program; if not, write to the Free 55 Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, 56 MA 02111-1307 USA 57 </para> 58 59 <para> 60 For more details see the file COPYING in the source 61 distribution of Linux. 62 </para> 63 </legalnotice> 64 </bookinfo> 65 66<toc></toc> 67 68 <chapter id="intro"> 69 <title>Introduction</title> 70 <para> 71 Linux provides an API which abstracts performing IO across all busses 72 and devices, allowing device drivers to be written independently of 73 bus type. 74 </para> 75 </chapter> 76 77 <chapter id="bugs"> 78 <title>Known Bugs And Assumptions</title> 79 <para> 80 None. 81 </para> 82 </chapter> 83 84 <chapter id="mmio"> 85 <title>Memory Mapped IO</title> 86 <sect1> 87 <title>Getting Access to the Device</title> 88 <para> 89 The most widely supported form of IO is memory mapped IO. 90 That is, a part of the CPU's address space is interpreted 91 not as accesses to memory, but as accesses to a device. Some 92 architectures define devices to be at a fixed address, but most 93 have some method of discovering devices. The PCI bus walk is a 94 good example of such a scheme. This document does not cover how 95 to receive such an address, but assumes you are starting with one. 96 Physical addresses are of type unsigned long. 97 </para> 98 99 <para> 100 This address should not be used directly. Instead, to get an 101 address suitable for passing to the accessor functions described 102 below, you should call <function>ioremap</function>. 103 An address suitable for accessing the device will be returned to you. 104 </para> 105 106 <para> 107 After you've finished using the device (say, in your module's 108 exit routine), call <function>iounmap</function> in order to return 109 the address space to the kernel. Most architectures allocate new 110 address space each time you call <function>ioremap</function>, and 111 they can run out unless you call <function>iounmap</function>. 112 </para> 113 </sect1> 114 115 <sect1> 116 <title>Accessing the device</title> 117 <para> 118 The part of the interface most used by drivers is reading and 119 writing memory-mapped registers on the device. Linux provides 120 interfaces to read and write 8-bit, 16-bit, 32-bit and 64-bit 121 quantities. Due to a historical accident, these are named byte, 122 word, long and quad accesses. Both read and write accesses are 123 supported; there is no prefetch support at this time. 124 </para> 125 126 <para> 127 The functions are named <function>readb</function>, 128 <function>readw</function>, <function>readl</function>, 129 <function>readq</function>, <function>writeb</function>, 130 <function>writew</function>, <function>writel</function> and 131 <function>writeq</function>. 132 </para> 133 134 <para> 135 Some devices (such as framebuffers) would like to use larger 136 transfers than 8 bytes at a time. For these devices, the 137 <function>memcpy_toio</function>, <function>memcpy_fromio</function> 138 and <function>memset_io</function> functions are provided. 139 Do not use memset or memcpy on IO addresses; they 140 are not guaranteed to copy data in order. 141 </para> 142 143 <para> 144 The read and write functions are defined to be ordered. That is the 145 compiler is not permitted to reorder the I/O sequence. When the 146 ordering can be compiler optimised, you can use <function> 147 __readb</function> and friends to indicate the relaxed ordering. Use 148 this with care. The <function>rmb</function> provides a read memory 149 barrier. The <function>wmb</function> provides a write memory barrier. 150 </para> 151 152 <para> 153 While the basic functions are defined to be synchronous with respect 154 to each other and ordered with respect to each other the busses the 155 devices sit on may themselves have asynchronocity. In paticular many 156 authors are burned by the fact that PCI bus writes are posted 157 asynchronously. A driver author must issue a read from the same 158 device to ensure that writes have occurred in the specific cases the 159 author cares. This kind of property cannot be hidden from driver 160 writers in the API. 161 </para> 162 </sect1> 163 164 <sect1> 165 <title>ISA legacy functions</title> 166 <para> 167 On older kernels (2.2 and earlier) the ISA bus could be read or 168 written with these functions and without ioremap being used. This is 169 no longer true in Linux 2.4. A set of equivalent functions exist for 170 easy legacy driver porting. The functions available are prefixed 171 with 'isa_' and are <function>isa_readb</function>, 172 <function>isa_writeb</function>, <function>isa_readw</function>, 173 <function>isa_writew</function>, <function>isa_readl</function>, 174 <function>isa_writel</function>, <function>isa_memcpy_fromio</function> 175 and <function>isa_memcpy_toio</function> 176 </para> 177 <para> 178 These functions should not be used in new drivers, and will 179 eventually be going away. 180 </para> 181 </sect1> 182 183 </chapter> 184 185 <chapter> 186 <title>Port Space Accesses</title> 187 <sect1> 188 <title>Port Space Explained</title> 189 190 <para> 191 Another form of IO commonly supported is Port Space. This is a 192 range of addresses separate to the normal memory address space. 193 Access to these addresses is generally not as fast as accesses 194 to the memory mapped addresses, and it also has a potentially 195 smaller address space. 196 </para> 197 198 <para> 199 Unlike memory mapped IO, no preparation is required 200 to access port space. 201 </para> 202 203 </sect1> 204 <sect1> 205 <title>Accessing Port Space</title> 206 <para> 207 Accesses to this space are provided through a set of functions 208 which allow 8-bit, 16-bit and 32-bit accesses; also 209 known as byte, word and long. These functions are 210 <function>inb</function>, <function>inw</function>, 211 <function>inl</function>, <function>outb</function>, 212 <function>outw</function> and <function>outl</function>. 213 </para> 214 215 <para> 216 Some variants are provided for these functions. Some devices 217 require that accesses to their ports are slowed down. This 218 functionality is provided by appending a <function>_p</function> 219 to the end of the function. There are also equivalents to memcpy. 220 The <function>ins</function> and <function>outs</function> 221 functions copy bytes, words or longs to the given port. 222 </para> 223 </sect1> 224 225 </chapter> 226 227 <chapter id="pubfunctions"> 228 <title>Public Functions Provided</title> 229!Einclude/asm-i386/io.h 230 </chapter> 231 232</book> 233