1 /* $NetBSD: pte.h,v 1.16 2007/10/18 15:28:37 yamt Exp $ */ 2 3 /* 4 * Copyright (c) 2001 Wasabi Systems, Inc. 5 * All rights reserved. 6 * 7 * Written by Frank van der Linden for Wasabi Systems, Inc. 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions 11 * are met: 12 * 1. Redistributions of source code must retain the above copyright 13 * notice, this list of conditions and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 3. All advertising materials mentioning features or use of this software 18 * must display the following acknowledgement: 19 * This product includes software developed for the NetBSD Project by 20 * Wasabi Systems, Inc. 21 * 4. The name of Wasabi Systems, Inc. may not be used to endorse 22 * or promote products derived from this software without specific prior 23 * written permission. 24 * 25 * THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND 26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 27 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 28 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL WASABI SYSTEMS, INC 29 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 30 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 31 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 32 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 33 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 34 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 35 * POSSIBILITY OF SUCH DAMAGE. 36 */ 37 38 /* 39 * 40 * Copyright (c) 1997 Charles D. Cranor and Washington University. 41 * All rights reserved. 42 * 43 * Redistribution and use in source and binary forms, with or without 44 * modification, are permitted provided that the following conditions 45 * are met: 46 * 1. Redistributions of source code must retain the above copyright 47 * notice, this list of conditions and the following disclaimer. 48 * 2. Redistributions in binary form must reproduce the above copyright 49 * notice, this list of conditions and the following disclaimer in the 50 * documentation and/or other materials provided with the distribution. 51 * 3. All advertising materials mentioning features or use of this software 52 * must display the following acknowledgment: 53 * This product includes software developed by Charles D. Cranor and 54 * Washington University. 55 * 4. The name of the author may not be used to endorse or promote products 56 * derived from this software without specific prior written permission. 57 * 58 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR 59 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 60 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 61 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, 62 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 63 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 64 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 65 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 66 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 67 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 68 */ 69 70 /* 71 * pte.h rewritten by chuck based on the jolitz version, plus random 72 * info on the pentium and other processors found on the net. the 73 * goal of this rewrite is to provide enough documentation on the MMU 74 * hardware that the reader will be able to understand it without having 75 * to refer to a hardware manual. 76 */ 77 78 #ifndef _I386_PTE_H_ 79 #define _I386_PTE_H_ 80 81 /* 82 * i386 MMU hardware structure: 83 * 84 * the i386 MMU is a two-level MMU which maps 4GB of virtual memory. 85 * the pagesize is 4K (4096 [0x1000] bytes), although newer pentium 86 * processors can support a 4MB pagesize as well. 87 * 88 * the first level table (segment table?) is called a "page directory" 89 * and it contains 1024 page directory entries (PDEs). each PDE is 90 * 4 bytes (an int), so a PD fits in a single 4K page. this page is 91 * the page directory page (PDP). each PDE in a PDP maps 4MB of space 92 * (1024 * 4MB = 4GB). a PDE contains the physical address of the 93 * second level table: the page table. or, if 4MB pages are being used, 94 * then the PDE contains the PA of the 4MB page being mapped. 95 * 96 * a page table consists of 1024 page table entries (PTEs). each PTE is 97 * 4 bytes (an int), so a page table also fits in a single 4K page. a 98 * 4K page being used as a page table is called a page table page (PTP). 99 * each PTE in a PTP maps one 4K page (1024 * 4K = 4MB). a PTE contains 100 * the physical address of the page it maps and some flag bits (described 101 * below). 102 * 103 * the processor has a special register, "cr3", which points to the 104 * the PDP which is currently controlling the mappings of the virtual 105 * address space. 106 * 107 * the following picture shows the translation process for a 4K page: 108 * 109 * %cr3 register [PA of PDP] 110 * | 111 * | 112 * | bits <31-22> of VA bits <21-12> of VA bits <11-0> 113 * | index the PDP (0 - 1023) index the PTP are the page offset 114 * | | | | 115 * | v | | 116 * +--->+----------+ | | 117 * | PD Page | PA of v | 118 * | |---PTP-------->+------------+ | 119 * | 1024 PDE | | page table |--PTE--+ | 120 * | entries | | (aka PTP) | | | 121 * +----------+ | 1024 PTE | | | 122 * | entries | | | 123 * +------------+ | | 124 * | | 125 * bits <31-12> bits <11-0> 126 * p h y s i c a l a d d r 127 * 128 * the i386 caches PTEs in a TLB. it is important to flush out old 129 * TLB mappings when making a change to a mappings. writing to the 130 * %cr3 will flush the entire TLB. newer processors also have an 131 * instruction that will invalidate the mapping of a single page (which 132 * is useful if you are changing a single mappings because it preserves 133 * all the cached TLB entries). 134 * 135 * as shows, bits 31-12 of the PTE contain PA of the page being mapped. 136 * the rest of the PTE is defined as follows: 137 * bit# name use 138 * 11 n/a available for OS use, hardware ignores it 139 * 10 n/a available for OS use, hardware ignores it 140 * 9 n/a available for OS use, hardware ignores it 141 * 8 G global bit (see discussion below) 142 * 7 PS page size [for PDEs] (0=4k, 1=4M <if supported>) 143 * 6 D dirty (modified) page 144 * 5 A accessed (referenced) page 145 * 4 PCD cache disable 146 * 3 PWT prevent write through (cache) 147 * 2 U/S user/supervisor bit (0=supervisor only, 1=both u&s) 148 * 1 R/W read/write bit (0=read only, 1=read-write) 149 * 0 P present (valid) 150 * 151 * notes: 152 * - on the i386 the R/W bit is ignored if processor is in supervisor 153 * state (bug!) 154 * - PS is only supported on newer processors 155 * - PTEs with the G bit are global in the sense that they are not 156 * flushed from the TLB when %cr3 is written (to flush, use the 157 * "flush single page" instruction). this is only supported on 158 * newer processors. this bit can be used to keep the kernel's 159 * TLB entries around while context switching. since the kernel 160 * is mapped into all processes at the same place it does not make 161 * sense to flush these entries when switching from one process' 162 * pmap to another. 163 */ 164 165 #if !defined(_LOCORE) 166 167 /* 168 * here we define the data types for PDEs and PTEs 169 */ 170 171 typedef uint32_t pd_entry_t; /* PDE */ 172 typedef uint32_t pt_entry_t; /* PTE */ 173 174 #endif 175 176 /* 177 * now we define various for playing with virtual addresses 178 */ 179 180 #define L1_SHIFT 12 181 #define L2_SHIFT 22 182 #define NBPD_L1 (1ULL << L1_SHIFT) /* # bytes mapped by L1 ent (4K) */ 183 #define NBPD_L2 (1ULL << L2_SHIFT) /* # bytes mapped by L2 ent (4MB) */ 184 185 #define L2_MASK 0xffc00000 186 #define L1_MASK 0x003ff000 187 188 #define L2_FRAME (L2_MASK) 189 #define L1_FRAME (L2_FRAME|L1_MASK) 190 191 /* 192 * here we define the bits of the PDE/PTE, as described above: 193 * 194 * XXXCDC: need to rename these (PG_u == ugly). 195 */ 196 197 #define PG_V 0x00000001 /* valid entry */ 198 #define PG_RO 0x00000000 /* read-only page */ 199 #define PG_RW 0x00000002 /* read-write page */ 200 #define PG_u 0x00000004 /* user accessible page */ 201 #define PG_PROT 0x00000806 /* all protection bits */ 202 #define PG_N 0x00000018 /* non-cacheable */ 203 #define PG_U 0x00000020 /* has been used */ 204 #define PG_M 0x00000040 /* has been modified */ 205 #define PG_PS 0x00000080 /* 4MB page size */ 206 #define PG_G 0x00000100 /* global, don't TLB flush */ 207 #define PG_AVAIL1 0x00000200 /* ignored by hardware */ 208 #define PG_AVAIL2 0x00000400 /* ignored by hardware */ 209 #define PG_AVAIL3 0x00000800 /* ignored by hardware */ 210 #define PG_FRAME 0xfffff000 /* page frame mask */ 211 212 #define PG_LGFRAME 0xffc00000 /* large (4MB) page frame mask */ 213 214 /* 215 * various short-hand protection codes 216 */ 217 218 #define PG_KR 0x00000000 /* kernel read-only */ 219 #define PG_KW 0x00000002 /* kernel read-write */ 220 #define PG_NX 0 /* dummy */ 221 222 /* 223 * page protection exception bits 224 */ 225 226 #define PGEX_P 0x01 /* protection violation (vs. no mapping) */ 227 #define PGEX_W 0x02 /* exception during a write cycle */ 228 #define PGEX_U 0x04 /* exception while in user mode (upl) */ 229 230 #endif /* _I386_PTE_H_ */ 231