Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 1 | THE LINUX/I386 BOOT PROTOCOL |
| 2 | ---------------------------- |
| 3 | |
| 4 | H. Peter Anvin <hpa@zytor.com> |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 5 | Last update 2007-05-16 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 6 | |
| 7 | On the i386 platform, the Linux kernel uses a rather complicated boot |
| 8 | convention. This has evolved partially due to historical aspects, as |
| 9 | well as the desire in the early days to have the kernel itself be a |
| 10 | bootable image, the complicated PC memory model and due to changed |
| 11 | expectations in the PC industry caused by the effective demise of |
| 12 | real-mode DOS as a mainstream operating system. |
| 13 | |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 14 | Currently, the following versions of the Linux/i386 boot protocol exist. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 15 | |
| 16 | Old kernels: zImage/Image support only. Some very early kernels |
| 17 | may not even support a command line. |
| 18 | |
| 19 | Protocol 2.00: (Kernel 1.3.73) Added bzImage and initrd support, as |
| 20 | well as a formalized way to communicate between the |
| 21 | boot loader and the kernel. setup.S made relocatable, |
| 22 | although the traditional setup area still assumed |
| 23 | writable. |
| 24 | |
| 25 | Protocol 2.01: (Kernel 1.3.76) Added a heap overrun warning. |
| 26 | |
| 27 | Protocol 2.02: (Kernel 2.4.0-test3-pre3) New command line protocol. |
| 28 | Lower the conventional memory ceiling. No overwrite |
| 29 | of the traditional setup area, thus making booting |
| 30 | safe for systems which use the EBDA from SMM or 32-bit |
| 31 | BIOS entry points. zImage deprecated but still |
| 32 | supported. |
| 33 | |
| 34 | Protocol 2.03: (Kernel 2.4.18-pre1) Explicitly makes the highest possible |
| 35 | initrd address available to the bootloader. |
| 36 | |
H. Peter Anvin | f8eeaaf | 2005-09-06 15:17:24 -0700 | [diff] [blame] | 37 | Protocol 2.04: (Kernel 2.6.14) Extend the syssize field to four bytes. |
Bernhard Walle | 8f9aeca | 2007-05-02 19:27:10 +0200 | [diff] [blame] | 38 | |
Vivek Goyal | be274ee | 2006-12-07 02:14:04 +0100 | [diff] [blame] | 39 | Protocol 2.05: (Kernel 2.6.20) Make protected mode kernel relocatable. |
| 40 | Introduce relocatable_kernel and kernel_alignment fields. |
H. Peter Anvin | f8eeaaf | 2005-09-06 15:17:24 -0700 | [diff] [blame] | 41 | |
Bernhard Walle | 8f9aeca | 2007-05-02 19:27:10 +0200 | [diff] [blame] | 42 | Protocol 2.06: (Kernel 2.6.22) Added a field that contains the size of |
| 43 | the boot command line |
| 44 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 45 | |
| 46 | **** MEMORY LAYOUT |
| 47 | |
| 48 | The traditional memory map for the kernel loader, used for Image or |
| 49 | zImage kernels, typically looks like: |
| 50 | |
| 51 | | | |
| 52 | 0A0000 +------------------------+ |
| 53 | | Reserved for BIOS | Do not use. Reserved for BIOS EBDA. |
| 54 | 09A000 +------------------------+ |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 55 | | Command line | |
| 56 | | Stack/heap | For use by the kernel real-mode code. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 57 | 098000 +------------------------+ |
| 58 | | Kernel setup | The kernel real-mode code. |
| 59 | 090200 +------------------------+ |
| 60 | | Kernel boot sector | The kernel legacy boot sector. |
| 61 | 090000 +------------------------+ |
| 62 | | Protected-mode kernel | The bulk of the kernel image. |
| 63 | 010000 +------------------------+ |
| 64 | | Boot loader | <- Boot sector entry point 0000:7C00 |
| 65 | 001000 +------------------------+ |
| 66 | | Reserved for MBR/BIOS | |
| 67 | 000800 +------------------------+ |
| 68 | | Typically used by MBR | |
| 69 | 000600 +------------------------+ |
| 70 | | BIOS use only | |
| 71 | 000000 +------------------------+ |
| 72 | |
| 73 | |
| 74 | When using bzImage, the protected-mode kernel was relocated to |
| 75 | 0x100000 ("high memory"), and the kernel real-mode block (boot sector, |
| 76 | setup, and stack/heap) was made relocatable to any address between |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 77 | 0x10000 and end of low memory. Unfortunately, in protocols 2.00 and |
| 78 | 2.01 the 0x90000+ memory range is still used internally by the kernel; |
| 79 | the 2.02 protocol resolves that problem. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 80 | |
| 81 | It is desirable to keep the "memory ceiling" -- the highest point in |
| 82 | low memory touched by the boot loader -- as low as possible, since |
| 83 | some newer BIOSes have begun to allocate some rather large amounts of |
| 84 | memory, called the Extended BIOS Data Area, near the top of low |
| 85 | memory. The boot loader should use the "INT 12h" BIOS call to verify |
| 86 | how much low memory is available. |
| 87 | |
| 88 | Unfortunately, if INT 12h reports that the amount of memory is too |
| 89 | low, there is usually nothing the boot loader can do but to report an |
| 90 | error to the user. The boot loader should therefore be designed to |
| 91 | take up as little space in low memory as it reasonably can. For |
| 92 | zImage or old bzImage kernels, which need data written into the |
| 93 | 0x90000 segment, the boot loader should make sure not to use memory |
| 94 | above the 0x9A000 point; too many BIOSes will break above that point. |
| 95 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 96 | For a modern bzImage kernel with boot protocol version >= 2.02, a |
| 97 | memory layout like the following is suggested: |
| 98 | |
| 99 | ~ ~ |
| 100 | | Protected-mode kernel | |
| 101 | 100000 +------------------------+ |
| 102 | | I/O memory hole | |
| 103 | 0A0000 +------------------------+ |
| 104 | | Reserved for BIOS | Leave as much as possible unused |
| 105 | ~ ~ |
| 106 | | Command line | (Can also be below the X+10000 mark) |
| 107 | X+10000 +------------------------+ |
| 108 | | Stack/heap | For use by the kernel real-mode code. |
| 109 | X+08000 +------------------------+ |
| 110 | | Kernel setup | The kernel real-mode code. |
| 111 | | Kernel boot sector | The kernel legacy boot sector. |
| 112 | X +------------------------+ |
| 113 | | Boot loader | <- Boot sector entry point 0000:7C00 |
| 114 | 001000 +------------------------+ |
| 115 | | Reserved for MBR/BIOS | |
| 116 | 000800 +------------------------+ |
| 117 | | Typically used by MBR | |
| 118 | 000600 +------------------------+ |
| 119 | | BIOS use only | |
| 120 | 000000 +------------------------+ |
| 121 | |
| 122 | ... where the address X is as low as the design of the boot loader |
| 123 | permits. |
| 124 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 125 | |
| 126 | **** THE REAL-MODE KERNEL HEADER |
| 127 | |
| 128 | In the following text, and anywhere in the kernel boot sequence, "a |
| 129 | sector" refers to 512 bytes. It is independent of the actual sector |
| 130 | size of the underlying medium. |
| 131 | |
| 132 | The first step in loading a Linux kernel should be to load the |
| 133 | real-mode code (boot sector and setup code) and then examine the |
| 134 | following header at offset 0x01f1. The real-mode code can total up to |
| 135 | 32K, although the boot loader may choose to load only the first two |
| 136 | sectors (1K) and then examine the bootup sector size. |
| 137 | |
| 138 | The header looks like: |
| 139 | |
| 140 | Offset Proto Name Meaning |
| 141 | /Size |
| 142 | |
H. Peter Anvin | f8eeaaf | 2005-09-06 15:17:24 -0700 | [diff] [blame] | 143 | 01F1/1 ALL(1 setup_sects The size of the setup in sectors |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 144 | 01F2/2 ALL root_flags If set, the root is mounted readonly |
H. Peter Anvin | f8eeaaf | 2005-09-06 15:17:24 -0700 | [diff] [blame] | 145 | 01F4/4 2.04+(2 syssize The size of the 32-bit code in 16-byte paras |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 146 | 01F8/2 ALL ram_size DO NOT USE - for bootsect.S use only |
| 147 | 01FA/2 ALL vid_mode Video mode control |
| 148 | 01FC/2 ALL root_dev Default root device number |
| 149 | 01FE/2 ALL boot_flag 0xAA55 magic number |
| 150 | 0200/2 2.00+ jump Jump instruction |
| 151 | 0202/4 2.00+ header Magic signature "HdrS" |
| 152 | 0206/2 2.00+ version Boot protocol version supported |
| 153 | 0208/4 2.00+ realmode_swtch Boot loader hook (see below) |
| 154 | 020C/2 2.00+ start_sys The load-low segment (0x1000) (obsolete) |
| 155 | 020E/2 2.00+ kernel_version Pointer to kernel version string |
| 156 | 0210/1 2.00+ type_of_loader Boot loader identifier |
| 157 | 0211/1 2.00+ loadflags Boot protocol option flags |
| 158 | 0212/2 2.00+ setup_move_size Move to high memory size (used with hooks) |
| 159 | 0214/4 2.00+ code32_start Boot loader hook (see below) |
| 160 | 0218/4 2.00+ ramdisk_image initrd load address (set by boot loader) |
| 161 | 021C/4 2.00+ ramdisk_size initrd size (set by boot loader) |
| 162 | 0220/4 2.00+ bootsect_kludge DO NOT USE - for bootsect.S use only |
| 163 | 0224/2 2.01+ heap_end_ptr Free memory after setup end |
| 164 | 0226/2 N/A pad1 Unused |
| 165 | 0228/4 2.02+ cmd_line_ptr 32-bit pointer to the kernel command line |
| 166 | 022C/4 2.03+ initrd_addr_max Highest legal initrd address |
Vivek Goyal | d263b21 | 2006-12-07 02:14:13 +0100 | [diff] [blame] | 167 | 0230/4 2.05+ kernel_alignment Physical addr alignment required for kernel |
| 168 | 0234/1 2.05+ relocatable_kernel Whether kernel is relocatable or not |
Bernhard Walle | 8f9aeca | 2007-05-02 19:27:10 +0200 | [diff] [blame] | 169 | 0235/3 N/A pad2 Unused |
| 170 | 0238/4 2.06+ cmdline_size Maximum size of the kernel command line |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 171 | |
H. Peter Anvin | f8eeaaf | 2005-09-06 15:17:24 -0700 | [diff] [blame] | 172 | (1) For backwards compatibility, if the setup_sects field contains 0, the |
| 173 | real value is 4. |
| 174 | |
| 175 | (2) For boot protocol prior to 2.04, the upper two bytes of the syssize |
| 176 | field are unusable, which means the size of a bzImage kernel |
| 177 | cannot be determined. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 178 | |
| 179 | If the "HdrS" (0x53726448) magic number is not found at offset 0x202, |
| 180 | the boot protocol version is "old". Loading an old kernel, the |
| 181 | following parameters should be assumed: |
| 182 | |
| 183 | Image type = zImage |
| 184 | initrd not supported |
| 185 | Real-mode kernel must be located at 0x90000. |
| 186 | |
| 187 | Otherwise, the "version" field contains the protocol version, |
| 188 | e.g. protocol version 2.01 will contain 0x0201 in this field. When |
| 189 | setting fields in the header, you must make sure only to set fields |
| 190 | supported by the protocol version in use. |
| 191 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 192 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 193 | **** DETAILS OF HEADER FIELDS |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 194 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 195 | For each field, some are information from the kernel to the bootloader |
| 196 | ("read"), some are expected to be filled out by the bootloader |
| 197 | ("write"), and some are expected to be read and modified by the |
| 198 | bootloader ("modify"). |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 199 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 200 | All general purpose boot loaders should write the fields marked |
| 201 | (obligatory). Boot loaders who want to load the kernel at a |
| 202 | nonstandard address should fill in the fields marked (reloc); other |
| 203 | boot loaders can ignore those fields. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 204 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 205 | Field name: setup_secs |
| 206 | Type: read |
| 207 | Offset/size: 0x1f1/1 |
| 208 | Protocol: ALL |
| 209 | |
| 210 | The size of the setup code in 512-byte sectors. If this field is |
| 211 | 0, the real value is 4. The real-mode code consists of the boot |
| 212 | sector (always one 512-byte sector) plus the setup code. |
| 213 | |
| 214 | Field name: root_flags |
| 215 | Type: modify (optional) |
| 216 | Offset/size: 0x1f2/2 |
| 217 | Protocol: ALL |
| 218 | |
| 219 | If this field is nonzero, the root defaults to readonly. The use of |
| 220 | this field is deprecated; use the "ro" or "rw" options on the |
| 221 | command line instead. |
| 222 | |
| 223 | Field name: syssize |
| 224 | Type: read |
| 225 | Offset/size: 0x1f4/4 (protocol 2.04+) 0x1f4/2 (protocol ALL) |
| 226 | Protocol: 2.04+ |
| 227 | |
| 228 | The size of the protected-mode code in units of 16-byte paragraphs. |
| 229 | For protocol versions older than 2.04 this field is only two bytes |
| 230 | wide, and therefore cannot be trusted for the size of a kernel if |
| 231 | the LOAD_HIGH flag is set. |
| 232 | |
| 233 | Field name: ram_size |
| 234 | Type: kernel internal |
| 235 | Offset/size: 0x1f8/2 |
| 236 | Protocol: ALL |
| 237 | |
| 238 | This field is obsolete. |
| 239 | |
| 240 | Field name: vid_mode |
| 241 | Type: modify (obligatory) |
| 242 | Offset/size: 0x1fa/2 |
| 243 | |
| 244 | Please see the section on SPECIAL COMMAND LINE OPTIONS. |
| 245 | |
| 246 | Field name: root_dev |
| 247 | Type: modify (optional) |
| 248 | Offset/size: 0x1fc/2 |
| 249 | Protocol: ALL |
| 250 | |
| 251 | The default root device device number. The use of this field is |
| 252 | deprecated, use the "root=" option on the command line instead. |
| 253 | |
| 254 | Field name: boot_flag |
| 255 | Type: read |
| 256 | Offset/size: 0x1fe/2 |
| 257 | Protocol: ALL |
| 258 | |
| 259 | Contains 0xAA55. This is the closest thing old Linux kernels have |
| 260 | to a magic number. |
| 261 | |
| 262 | Field name: jump |
| 263 | Type: read |
| 264 | Offset/size: 0x200/2 |
| 265 | Protocol: 2.00+ |
| 266 | |
| 267 | Contains an x86 jump instruction, 0xEB followed by a signed offset |
| 268 | relative to byte 0x202. This can be used to determine the size of |
| 269 | the header. |
| 270 | |
| 271 | Field name: header |
| 272 | Type: read |
| 273 | Offset/size: 0x202/4 |
| 274 | Protocol: 2.00+ |
| 275 | |
| 276 | Contains the magic number "HdrS" (0x53726448). |
| 277 | |
| 278 | Field name: version |
| 279 | Type: read |
| 280 | Offset/size: 0x206/2 |
| 281 | Protocol: 2.00+ |
| 282 | |
| 283 | Contains the boot protocol version, e.g. 0x0204 for version 2.04. |
| 284 | |
| 285 | Field name: readmode_swtch |
| 286 | Type: modify (optional) |
| 287 | Offset/size: 0x208/4 |
| 288 | Protocol: 2.00+ |
| 289 | |
| 290 | Boot loader hook (see separate chapter.) |
| 291 | |
| 292 | Field name: start_sys |
| 293 | Type: read |
| 294 | Offset/size: 0x20c/4 |
| 295 | Protocol: 2.00+ |
| 296 | |
| 297 | The load low segment (0x1000). Obsolete. |
| 298 | |
| 299 | Field name: kernel_version |
| 300 | Type: read |
| 301 | Offset/size: 0x20e/2 |
| 302 | Protocol: 2.00+ |
| 303 | |
| 304 | If set to a nonzero value, contains a pointer to a NUL-terminated |
| 305 | human-readable kernel version number string, less 0x200. This can |
| 306 | be used to display the kernel version to the user. This value |
| 307 | should be less than (0x200*setup_sects). For example, if this value |
| 308 | is set to 0x1c00, the kernel version number string can be found at |
| 309 | offset 0x1e00 in the kernel file. This is a valid value if and only |
| 310 | if the "setup_sects" field contains the value 14 or higher. |
| 311 | |
| 312 | Field name: type_of_loader |
| 313 | Type: write (obligatory) |
| 314 | Offset/size: 0x210/1 |
| 315 | Protocol: 2.00+ |
| 316 | |
| 317 | If your boot loader has an assigned id (see table below), enter |
| 318 | 0xTV here, where T is an identifier for the boot loader and V is |
| 319 | a version number. Otherwise, enter 0xFF here. |
| 320 | |
| 321 | Assigned boot loader ids: |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 322 | 0 LILO (0x00 reserved for pre-2.00 bootloader) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 323 | 1 Loadlin |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 324 | 2 bootsect-loader (0x20, all other values reserved) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 325 | 3 SYSLINUX |
| 326 | 4 EtherBoot |
| 327 | 5 ELILO |
| 328 | 7 GRuB |
| 329 | 8 U-BOOT |
Jeremy Fitzhardinge | 354332e | 2006-09-12 20:35:57 -0700 | [diff] [blame] | 330 | 9 Xen |
H. Peter Anvin | c229ec5 | 2007-01-26 09:15:47 -0800 | [diff] [blame] | 331 | A Gujin |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 332 | B Qemu |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 333 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 334 | Please contact <hpa@zytor.com> if you need a bootloader ID |
| 335 | value assigned. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 336 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 337 | Field name: loadflags |
| 338 | Type: modify (obligatory) |
| 339 | Offset/size: 0x211/1 |
| 340 | Protocol: 2.00+ |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 341 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 342 | This field is a bitmask. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 343 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 344 | Bit 0 (read): LOADED_HIGH |
| 345 | - If 0, the protected-mode code is loaded at 0x10000. |
| 346 | - If 1, the protected-mode code is loaded at 0x100000. |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 347 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 348 | Bit 7 (write): CAN_USE_HEAP |
| 349 | Set this bit to 1 to indicate that the value entered in the |
| 350 | heap_end_ptr is valid. If this field is clear, some setup code |
| 351 | functionality will be disabled. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 352 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 353 | Field name: setup_move_size |
| 354 | Type: modify (obligatory) |
| 355 | Offset/size: 0x212/2 |
| 356 | Protocol: 2.00-2.01 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 357 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 358 | When using protocol 2.00 or 2.01, if the real mode kernel is not |
| 359 | loaded at 0x90000, it gets moved there later in the loading |
| 360 | sequence. Fill in this field if you want additional data (such as |
| 361 | the kernel command line) moved in addition to the real-mode kernel |
| 362 | itself. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 363 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 364 | The unit is bytes starting with the beginning of the boot sector. |
| 365 | |
| 366 | This field is can be ignored when the protocol is 2.02 or higher, or |
| 367 | if the real-mode code is loaded at 0x90000. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 368 | |
H. Peter Anvin | dec04cf | 2007-05-17 15:50:47 -0700 | [diff] [blame^] | 369 | Field name: code32_start |
| 370 | Type: modify (optional, reloc) |
| 371 | Offset/size: 0x214/4 |
| 372 | Protocol: 2.00+ |
| 373 | |
| 374 | The address to jump to in protected mode. This defaults to the load |
| 375 | address of the kernel, and can be used by the boot loader to |
| 376 | determine the proper load address. |
| 377 | |
| 378 | This field can be modified for two purposes: |
| 379 | |
| 380 | 1. as a boot loader hook (see separate chapter.) |
| 381 | |
| 382 | 2. if a bootloader which does not install a hook loads a |
| 383 | relocatable kernel at a nonstandard address it will have to modify |
| 384 | this field to point to the load address. |
| 385 | |
| 386 | Field name: ramdisk_image |
| 387 | Type: write (obligatory) |
| 388 | Offset/size: 0x218/4 |
| 389 | Protocol: 2.00+ |
| 390 | |
| 391 | The 32-bit linear address of the initial ramdisk or ramfs. Leave at |
| 392 | zero if there is no initial ramdisk/ramfs. |
| 393 | |
| 394 | Field name: ramdisk_size |
| 395 | Type: write (obligatory) |
| 396 | Offset/size: 0x21c/4 |
| 397 | Protocol: 2.00+ |
| 398 | |
| 399 | Size of the initial ramdisk or ramfs. Leave at zero if there is no |
| 400 | initial ramdisk/ramfs. |
| 401 | |
| 402 | Field name: bootsect_kludge |
| 403 | Type: kernel internal |
| 404 | Offset/size: 0x220/4 |
| 405 | Protocol: 2.00+ |
| 406 | |
| 407 | This field is obsolete. |
| 408 | |
| 409 | Field name: heap_end_ptr |
| 410 | Type: write (obligatory) |
| 411 | Offset/size: 0x224/2 |
| 412 | Protocol: 2.01+ |
| 413 | |
| 414 | Set this field to the offset (from the beginning of the real-mode |
| 415 | code) of the end of the setup stack/heap, minus 0x0200. |
| 416 | |
| 417 | Field name: cmd_line_ptr |
| 418 | Type: write (obligatory) |
| 419 | Offset/size: 0x228/4 |
| 420 | Protocol: 2.02+ |
| 421 | |
| 422 | Set this field to the linear address of the kernel command line. |
| 423 | The kernel command line can be located anywhere between the end of |
| 424 | the setup heap and 0xA0000; it does not have to be located in the |
| 425 | same 64K segment as the real-mode code itself. |
| 426 | |
| 427 | Fill in this field even if your boot loader does not support a |
| 428 | command line, in which case you can point this to an empty string |
| 429 | (or better yet, to the string "auto".) If this field is left at |
| 430 | zero, the kernel will assume that your boot loader does not support |
| 431 | the 2.02+ protocol. |
| 432 | |
| 433 | Field name: initrd_addr_max |
| 434 | Type: read |
| 435 | Offset/size: 0x22c/4 |
| 436 | Protocol: 2.03+ |
| 437 | |
| 438 | The maximum address that may be occupied by the initial |
| 439 | ramdisk/ramfs contents. For boot protocols 2.02 or earlier, this |
| 440 | field is not present, and the maximum address is 0x37FFFFFF. (This |
| 441 | address is defined as the address of the highest safe byte, so if |
| 442 | your ramdisk is exactly 131072 bytes long and this field is |
| 443 | 0x37FFFFFF, you can start your ramdisk at 0x37FE0000.) |
| 444 | |
| 445 | Field name: kernel_alignment |
| 446 | Type: read (reloc) |
| 447 | Offset/size: 0x230/4 |
| 448 | Protocol: 2.05+ |
| 449 | |
| 450 | Alignment unit required by the kernel (if relocatable_kernel is true.) |
| 451 | |
| 452 | Field name: relocatable_kernel |
| 453 | Type: read (reloc) |
| 454 | Offset/size: 0x234/1 |
| 455 | Protocol: 2.05+ |
| 456 | |
| 457 | If this field is nonzero, the protected-mode part of the kernel can |
| 458 | be loaded at any address that satisfies the kernel_alignment field. |
| 459 | After loading, the boot loader must set the code32_start field to |
| 460 | point to the loaded code, or to a boot loader hook. |
| 461 | |
| 462 | Field name: cmdline_size |
| 463 | Type: read |
| 464 | Offset/size: 0x238/4 |
| 465 | Protocol: 2.06+ |
| 466 | |
| 467 | The maximum size of the command line without the terminating |
| 468 | zero. This means that the command line can contain at most |
| 469 | cmdline_size characters. With protocol version 2.05 and earlier, the |
| 470 | maximum size was 255. |
Bernhard Walle | 8f9aeca | 2007-05-02 19:27:10 +0200 | [diff] [blame] | 471 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 472 | |
| 473 | **** THE KERNEL COMMAND LINE |
| 474 | |
| 475 | The kernel command line has become an important way for the boot |
| 476 | loader to communicate with the kernel. Some of its options are also |
| 477 | relevant to the boot loader itself, see "special command line options" |
| 478 | below. |
| 479 | |
Bernhard Walle | 8f9aeca | 2007-05-02 19:27:10 +0200 | [diff] [blame] | 480 | The kernel command line is a null-terminated string. The maximum |
| 481 | length can be retrieved from the field cmdline_size. Before protocol |
| 482 | version 2.06, the maximum was 255 characters. A string that is too |
| 483 | long will be automatically truncated by the kernel. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 484 | |
| 485 | If the boot protocol version is 2.02 or later, the address of the |
| 486 | kernel command line is given by the header field cmd_line_ptr (see |
H. Peter Anvin | f8eeaaf | 2005-09-06 15:17:24 -0700 | [diff] [blame] | 487 | above.) This address can be anywhere between the end of the setup |
| 488 | heap and 0xA0000. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 489 | |
| 490 | If the protocol version is *not* 2.02 or higher, the kernel |
| 491 | command line is entered using the following protocol: |
| 492 | |
| 493 | At offset 0x0020 (word), "cmd_line_magic", enter the magic |
| 494 | number 0xA33F. |
| 495 | |
| 496 | At offset 0x0022 (word), "cmd_line_offset", enter the offset |
| 497 | of the kernel command line (relative to the start of the |
| 498 | real-mode kernel). |
| 499 | |
| 500 | The kernel command line *must* be within the memory region |
| 501 | covered by setup_move_size, so you may need to adjust this |
| 502 | field. |
| 503 | |
| 504 | |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 505 | **** MEMORY LAYOUT OF THE REAL-MODE CODE |
| 506 | |
| 507 | The real-mode code requires a stack/heap to be set up, as well as |
| 508 | memory allocated for the kernel command line. This needs to be done |
| 509 | in the real-mode accessible memory in bottom megabyte. |
| 510 | |
| 511 | It should be noted that modern machines often have a sizable Extended |
| 512 | BIOS Data Area (EBDA). As a result, it is advisable to use as little |
| 513 | of the low megabyte as possible. |
| 514 | |
| 515 | Unfortunately, under the following circumstances the 0x90000 memory |
| 516 | segment has to be used: |
| 517 | |
| 518 | - When loading a zImage kernel ((loadflags & 0x01) == 0). |
| 519 | - When loading a 2.01 or earlier boot protocol kernel. |
| 520 | |
| 521 | -> For the 2.00 and 2.01 boot protocols, the real-mode code |
| 522 | can be loaded at another address, but it is internally |
| 523 | relocated to 0x90000. For the "old" protocol, the |
| 524 | real-mode code must be loaded at 0x90000. |
| 525 | |
| 526 | When loading at 0x90000, avoid using memory above 0x9a000. |
| 527 | |
| 528 | For boot protocol 2.02 or higher, the command line does not have to be |
| 529 | located in the same 64K segment as the real-mode setup code; it is |
| 530 | thus permitted to give the stack/heap the full 64K segment and locate |
| 531 | the command line above it. |
| 532 | |
| 533 | The kernel command line should not be located below the real-mode |
| 534 | code, nor should it be located in high memory. |
| 535 | |
| 536 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 537 | **** SAMPLE BOOT CONFIGURATION |
| 538 | |
| 539 | As a sample configuration, assume the following layout of the real |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 540 | mode segment: |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 541 | |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 542 | When loading below 0x90000, use the entire segment: |
| 543 | |
| 544 | 0x0000-0x7fff Real mode kernel |
| 545 | 0x8000-0xdfff Stack and heap |
| 546 | 0xe000-0xffff Kernel command line |
| 547 | |
| 548 | When loading at 0x90000 OR the protocol version is 2.01 or earlier: |
| 549 | |
| 550 | 0x0000-0x7fff Real mode kernel |
| 551 | 0x8000-0x97ff Stack and heap |
| 552 | 0x9800-0x9fff Kernel command line |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 553 | |
| 554 | Such a boot loader should enter the following fields in the header: |
| 555 | |
| 556 | unsigned long base_ptr; /* base address for real-mode segment */ |
| 557 | |
| 558 | if ( setup_sects == 0 ) { |
| 559 | setup_sects = 4; |
| 560 | } |
| 561 | |
| 562 | if ( protocol >= 0x0200 ) { |
| 563 | type_of_loader = <type code>; |
| 564 | if ( loading_initrd ) { |
| 565 | ramdisk_image = <initrd_address>; |
| 566 | ramdisk_size = <initrd_size>; |
| 567 | } |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 568 | |
| 569 | if ( protocol >= 0x0202 && loadflags & 0x01 ) |
| 570 | heap_end = 0xe000; |
| 571 | else |
| 572 | heap_end = 0x9800; |
| 573 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 574 | if ( protocol >= 0x0201 ) { |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 575 | heap_end_ptr = heap_end - 0x200; |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 576 | loadflags |= 0x80; /* CAN_USE_HEAP */ |
| 577 | } |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 578 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 579 | if ( protocol >= 0x0202 ) { |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 580 | cmd_line_ptr = base_ptr + heap_end; |
| 581 | strcpy(cmd_line_ptr, cmdline); |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 582 | } else { |
| 583 | cmd_line_magic = 0xA33F; |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 584 | cmd_line_offset = heap_end; |
| 585 | setup_move_size = heap_end + strlen(cmdline)+1; |
| 586 | strcpy(base_ptr+cmd_line_offset, cmdline); |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 587 | } |
| 588 | } else { |
| 589 | /* Very old kernel */ |
| 590 | |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 591 | heap_end = 0x9800; |
| 592 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 593 | cmd_line_magic = 0xA33F; |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 594 | cmd_line_offset = heap_end; |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 595 | |
| 596 | /* A very old kernel MUST have its real-mode code |
| 597 | loaded at 0x90000 */ |
| 598 | |
| 599 | if ( base_ptr != 0x90000 ) { |
| 600 | /* Copy the real-mode kernel */ |
| 601 | memcpy(0x90000, base_ptr, (setup_sects+1)*512); |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 602 | base_ptr = 0x90000; /* Relocated */ |
| 603 | } |
| 604 | |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 605 | strcpy(0x90000+cmd_line_offset, cmdline); |
| 606 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 607 | /* It is recommended to clear memory up to the 32K mark */ |
| 608 | memset(0x90000 + (setup_sects+1)*512, 0, |
| 609 | (64-(setup_sects+1))*512); |
| 610 | } |
| 611 | |
| 612 | |
| 613 | **** LOADING THE REST OF THE KERNEL |
| 614 | |
H. Peter Anvin | f8eeaaf | 2005-09-06 15:17:24 -0700 | [diff] [blame] | 615 | The 32-bit (non-real-mode) kernel starts at offset (setup_sects+1)*512 |
| 616 | in the kernel file (again, if setup_sects == 0 the real value is 4.) |
| 617 | It should be loaded at address 0x10000 for Image/zImage kernels and |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 618 | 0x100000 for bzImage kernels. |
| 619 | |
| 620 | The kernel is a bzImage kernel if the protocol >= 2.00 and the 0x01 |
| 621 | bit (LOAD_HIGH) in the loadflags field is set: |
| 622 | |
| 623 | is_bzImage = (protocol >= 0x0200) && (loadflags & 0x01); |
| 624 | load_address = is_bzImage ? 0x100000 : 0x10000; |
| 625 | |
| 626 | Note that Image/zImage kernels can be up to 512K in size, and thus use |
| 627 | the entire 0x10000-0x90000 range of memory. This means it is pretty |
| 628 | much a requirement for these kernels to load the real-mode part at |
| 629 | 0x90000. bzImage kernels allow much more flexibility. |
| 630 | |
| 631 | |
| 632 | **** SPECIAL COMMAND LINE OPTIONS |
| 633 | |
| 634 | If the command line provided by the boot loader is entered by the |
| 635 | user, the user may expect the following command line options to work. |
| 636 | They should normally not be deleted from the kernel command line even |
| 637 | though not all of them are actually meaningful to the kernel. Boot |
| 638 | loader authors who need additional command line options for the boot |
| 639 | loader itself should get them registered in |
| 640 | Documentation/kernel-parameters.txt to make sure they will not |
| 641 | conflict with actual kernel options now or in the future. |
| 642 | |
| 643 | vga=<mode> |
| 644 | <mode> here is either an integer (in C notation, either |
| 645 | decimal, octal, or hexadecimal) or one of the strings |
| 646 | "normal" (meaning 0xFFFF), "ext" (meaning 0xFFFE) or "ask" |
| 647 | (meaning 0xFFFD). This value should be entered into the |
| 648 | vid_mode field, as it is used by the kernel before the command |
| 649 | line is parsed. |
| 650 | |
| 651 | mem=<size> |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 652 | <size> is an integer in C notation optionally followed by |
| 653 | (case insensitive) K, M, G, T, P or E (meaning << 10, << 20, |
| 654 | << 30, << 40, << 50 or << 60). This specifies the end of |
| 655 | memory to the kernel. This affects the possible placement of |
| 656 | an initrd, since an initrd should be placed near end of |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 657 | memory. Note that this is an option to *both* the kernel and |
| 658 | the bootloader! |
| 659 | |
| 660 | initrd=<file> |
| 661 | An initrd should be loaded. The meaning of <file> is |
| 662 | obviously bootloader-dependent, and some boot loaders |
| 663 | (e.g. LILO) do not have such a command. |
| 664 | |
| 665 | In addition, some boot loaders add the following options to the |
| 666 | user-specified command line: |
| 667 | |
| 668 | BOOT_IMAGE=<file> |
| 669 | The boot image which was loaded. Again, the meaning of <file> |
| 670 | is obviously bootloader-dependent. |
| 671 | |
| 672 | auto |
| 673 | The kernel was booted without explicit user intervention. |
| 674 | |
| 675 | If these options are added by the boot loader, it is highly |
| 676 | recommended that they are located *first*, before the user-specified |
| 677 | or configuration-specified command line. Otherwise, "init=/bin/sh" |
| 678 | gets confused by the "auto" option. |
| 679 | |
| 680 | |
| 681 | **** RUNNING THE KERNEL |
| 682 | |
| 683 | The kernel is started by jumping to the kernel entry point, which is |
| 684 | located at *segment* offset 0x20 from the start of the real mode |
| 685 | kernel. This means that if you loaded your real-mode kernel code at |
| 686 | 0x90000, the kernel entry point is 9020:0000. |
| 687 | |
| 688 | At entry, ds = es = ss should point to the start of the real-mode |
| 689 | kernel code (0x9000 if the code is loaded at 0x90000), sp should be |
| 690 | set up properly, normally pointing to the top of the heap, and |
| 691 | interrupts should be disabled. Furthermore, to guard against bugs in |
| 692 | the kernel, it is recommended that the boot loader sets fs = gs = ds = |
| 693 | es = ss. |
| 694 | |
| 695 | In our example from above, we would do: |
| 696 | |
| 697 | /* Note: in the case of the "old" kernel protocol, base_ptr must |
| 698 | be == 0x90000 at this point; see the previous sample code */ |
| 699 | |
| 700 | seg = base_ptr >> 4; |
| 701 | |
| 702 | cli(); /* Enter with interrupts disabled! */ |
| 703 | |
| 704 | /* Set up the real-mode kernel stack */ |
| 705 | _SS = seg; |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 706 | _SP = heap_end; |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 707 | |
| 708 | _DS = _ES = _FS = _GS = seg; |
| 709 | jmp_far(seg+0x20, 0); /* Run the kernel */ |
| 710 | |
| 711 | If your boot sector accesses a floppy drive, it is recommended to |
| 712 | switch off the floppy motor before running the kernel, since the |
| 713 | kernel boot leaves interrupts off and thus the motor will not be |
| 714 | switched off, especially if the loaded kernel has the floppy driver as |
| 715 | a demand-loaded module! |
| 716 | |
| 717 | |
| 718 | **** ADVANCED BOOT TIME HOOKS |
| 719 | |
| 720 | If the boot loader runs in a particularly hostile environment (such as |
| 721 | LOADLIN, which runs under DOS) it may be impossible to follow the |
| 722 | standard memory location requirements. Such a boot loader may use the |
| 723 | following hooks that, if set, are invoked by the kernel at the |
| 724 | appropriate time. The use of these hooks should probably be |
| 725 | considered an absolutely last resort! |
| 726 | |
| 727 | IMPORTANT: All the hooks are required to preserve %esp, %ebp, %esi and |
| 728 | %edi across invocation. |
| 729 | |
| 730 | realmode_swtch: |
| 731 | A 16-bit real mode far subroutine invoked immediately before |
| 732 | entering protected mode. The default routine disables NMI, so |
| 733 | your routine should probably do so, too. |
| 734 | |
| 735 | code32_start: |
| 736 | A 32-bit flat-mode routine *jumped* to immediately after the |
| 737 | transition to protected mode, but before the kernel is |
H. Peter Anvin | de372ec | 2007-05-08 20:37:02 -0700 | [diff] [blame] | 738 | uncompressed. No segments, except CS, are guaranteed to be |
| 739 | set up (current kernels do, but older ones do not); you should |
| 740 | set them up to BOOT_DS (0x18) yourself. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 741 | |
| 742 | After completing your hook, you should jump to the address |
| 743 | that was in this field before your boot loader overwrote it. |