0001 .. SPDX-License-Identifier: GPL-2.0
0002
0003 ===========================
0004 The Linux/x86 Boot Protocol
0005 ===========================
0006
0007 On the x86 platform, the Linux kernel uses a rather complicated boot
0008 convention. This has evolved partially due to historical aspects, as
0009 well as the desire in the early days to have the kernel itself be a
0010 bootable image, the complicated PC memory model and due to changed
0011 expectations in the PC industry caused by the effective demise of
0012 real-mode DOS as a mainstream operating system.
0013
0014 Currently, the following versions of the Linux/x86 boot protocol exist.
0015
0016 ============= ============================================================
0017 Old kernels zImage/Image support only. Some very early kernels
0018 may not even support a command line.
0019
0020 Protocol 2.00 (Kernel 1.3.73) Added bzImage and initrd support, as
0021 well as a formalized way to communicate between the
0022 boot loader and the kernel. setup.S made relocatable,
0023 although the traditional setup area still assumed
0024 writable.
0025
0026 Protocol 2.01 (Kernel 1.3.76) Added a heap overrun warning.
0027
0028 Protocol 2.02 (Kernel 2.4.0-test3-pre3) New command line protocol.
0029 Lower the conventional memory ceiling. No overwrite
0030 of the traditional setup area, thus making booting
0031 safe for systems which use the EBDA from SMM or 32-bit
0032 BIOS entry points. zImage deprecated but still
0033 supported.
0034
0035 Protocol 2.03 (Kernel 2.4.18-pre1) Explicitly makes the highest possible
0036 initrd address available to the bootloader.
0037
0038 Protocol 2.04 (Kernel 2.6.14) Extend the syssize field to four bytes.
0039
0040 Protocol 2.05 (Kernel 2.6.20) Make protected mode kernel relocatable.
0041 Introduce relocatable_kernel and kernel_alignment fields.
0042
0043 Protocol 2.06 (Kernel 2.6.22) Added a field that contains the size of
0044 the boot command line.
0045
0046 Protocol 2.07 (Kernel 2.6.24) Added paravirtualised boot protocol.
0047 Introduced hardware_subarch and hardware_subarch_data
0048 and KEEP_SEGMENTS flag in load_flags.
0049
0050 Protocol 2.08 (Kernel 2.6.26) Added crc32 checksum and ELF format
0051 payload. Introduced payload_offset and payload_length
0052 fields to aid in locating the payload.
0053
0054 Protocol 2.09 (Kernel 2.6.26) Added a field of 64-bit physical
0055 pointer to single linked list of struct setup_data.
0056
0057 Protocol 2.10 (Kernel 2.6.31) Added a protocol for relaxed alignment
0058 beyond the kernel_alignment added, new init_size and
0059 pref_address fields. Added extended boot loader IDs.
0060
0061 Protocol 2.11 (Kernel 3.6) Added a field for offset of EFI handover
0062 protocol entry point.
0063
0064 Protocol 2.12 (Kernel 3.8) Added the xloadflags field and extension fields
0065 to struct boot_params for loading bzImage and ramdisk
0066 above 4G in 64bit.
0067
0068 Protocol 2.13 (Kernel 3.14) Support 32- and 64-bit flags being set in
0069 xloadflags to support booting a 64-bit kernel from 32-bit
0070 EFI
0071
0072 Protocol 2.14 BURNT BY INCORRECT COMMIT
0073 ae7e1238e68f2a472a125673ab506d49158c1889
0074 (x86/boot: Add ACPI RSDP address to setup_header)
0075 DO NOT USE!!! ASSUME SAME AS 2.13.
0076
0077 Protocol 2.15 (Kernel 5.5) Added the kernel_info and kernel_info.setup_type_max.
0078 ============= ============================================================
0079
0080 .. note::
0081 The protocol version number should be changed only if the setup header
0082 is changed. There is no need to update the version number if boot_params
0083 or kernel_info are changed. Additionally, it is recommended to use
0084 xloadflags (in this case the protocol version number should not be
0085 updated either) or kernel_info to communicate supported Linux kernel
0086 features to the boot loader. Due to very limited space available in
0087 the original setup header every update to it should be considered
0088 with great care. Starting from the protocol 2.15 the primary way to
0089 communicate things to the boot loader is the kernel_info.
0090
0091
0092 Memory Layout
0093 =============
0094
0095 The traditional memory map for the kernel loader, used for Image or
0096 zImage kernels, typically looks like::
0097
0098 | |
0099 0A0000 +------------------------+
0100 | Reserved for BIOS | Do not use. Reserved for BIOS EBDA.
0101 09A000 +------------------------+
0102 | Command line |
0103 | Stack/heap | For use by the kernel real-mode code.
0104 098000 +------------------------+
0105 | Kernel setup | The kernel real-mode code.
0106 090200 +------------------------+
0107 | Kernel boot sector | The kernel legacy boot sector.
0108 090000 +------------------------+
0109 | Protected-mode kernel | The bulk of the kernel image.
0110 010000 +------------------------+
0111 | Boot loader | <- Boot sector entry point 0000:7C00
0112 001000 +------------------------+
0113 | Reserved for MBR/BIOS |
0114 000800 +------------------------+
0115 | Typically used by MBR |
0116 000600 +------------------------+
0117 | BIOS use only |
0118 000000 +------------------------+
0119
0120 When using bzImage, the protected-mode kernel was relocated to
0121 0x100000 ("high memory"), and the kernel real-mode block (boot sector,
0122 setup, and stack/heap) was made relocatable to any address between
0123 0x10000 and end of low memory. Unfortunately, in protocols 2.00 and
0124 2.01 the 0x90000+ memory range is still used internally by the kernel;
0125 the 2.02 protocol resolves that problem.
0126
0127 It is desirable to keep the "memory ceiling" -- the highest point in
0128 low memory touched by the boot loader -- as low as possible, since
0129 some newer BIOSes have begun to allocate some rather large amounts of
0130 memory, called the Extended BIOS Data Area, near the top of low
0131 memory. The boot loader should use the "INT 12h" BIOS call to verify
0132 how much low memory is available.
0133
0134 Unfortunately, if INT 12h reports that the amount of memory is too
0135 low, there is usually nothing the boot loader can do but to report an
0136 error to the user. The boot loader should therefore be designed to
0137 take up as little space in low memory as it reasonably can. For
0138 zImage or old bzImage kernels, which need data written into the
0139 0x90000 segment, the boot loader should make sure not to use memory
0140 above the 0x9A000 point; too many BIOSes will break above that point.
0141
0142 For a modern bzImage kernel with boot protocol version >= 2.02, a
0143 memory layout like the following is suggested::
0144
0145 ~ ~
0146 | Protected-mode kernel |
0147 100000 +------------------------+
0148 | I/O memory hole |
0149 0A0000 +------------------------+
0150 | Reserved for BIOS | Leave as much as possible unused
0151 ~ ~
0152 | Command line | (Can also be below the X+10000 mark)
0153 X+10000 +------------------------+
0154 | Stack/heap | For use by the kernel real-mode code.
0155 X+08000 +------------------------+
0156 | Kernel setup | The kernel real-mode code.
0157 | Kernel boot sector | The kernel legacy boot sector.
0158 X +------------------------+
0159 | Boot loader | <- Boot sector entry point 0000:7C00
0160 001000 +------------------------+
0161 | Reserved for MBR/BIOS |
0162 000800 +------------------------+
0163 | Typically used by MBR |
0164 000600 +------------------------+
0165 | BIOS use only |
0166 000000 +------------------------+
0167
0168 ... where the address X is as low as the design of the boot loader permits.
0169
0170
0171 The Real-Mode Kernel Header
0172 ===========================
0173
0174 In the following text, and anywhere in the kernel boot sequence, "a
0175 sector" refers to 512 bytes. It is independent of the actual sector
0176 size of the underlying medium.
0177
0178 The first step in loading a Linux kernel should be to load the
0179 real-mode code (boot sector and setup code) and then examine the
0180 following header at offset 0x01f1. The real-mode code can total up to
0181 32K, although the boot loader may choose to load only the first two
0182 sectors (1K) and then examine the bootup sector size.
0183
0184 The header looks like:
0185
0186 =========== ======== ===================== ============================================
0187 Offset/Size Proto Name Meaning
0188 =========== ======== ===================== ============================================
0189 01F1/1 ALL(1) setup_sects The size of the setup in sectors
0190 01F2/2 ALL root_flags If set, the root is mounted readonly
0191 01F4/4 2.04+(2) syssize The size of the 32-bit code in 16-byte paras
0192 01F8/2 ALL ram_size DO NOT USE - for bootsect.S use only
0193 01FA/2 ALL vid_mode Video mode control
0194 01FC/2 ALL root_dev Default root device number
0195 01FE/2 ALL boot_flag 0xAA55 magic number
0196 0200/2 2.00+ jump Jump instruction
0197 0202/4 2.00+ header Magic signature "HdrS"
0198 0206/2 2.00+ version Boot protocol version supported
0199 0208/4 2.00+ realmode_swtch Boot loader hook (see below)
0200 020C/2 2.00+ start_sys_seg The load-low segment (0x1000) (obsolete)
0201 020E/2 2.00+ kernel_version Pointer to kernel version string
0202 0210/1 2.00+ type_of_loader Boot loader identifier
0203 0211/1 2.00+ loadflags Boot protocol option flags
0204 0212/2 2.00+ setup_move_size Move to high memory size (used with hooks)
0205 0214/4 2.00+ code32_start Boot loader hook (see below)
0206 0218/4 2.00+ ramdisk_image initrd load address (set by boot loader)
0207 021C/4 2.00+ ramdisk_size initrd size (set by boot loader)
0208 0220/4 2.00+ bootsect_kludge DO NOT USE - for bootsect.S use only
0209 0224/2 2.01+ heap_end_ptr Free memory after setup end
0210 0226/1 2.02+(3) ext_loader_ver Extended boot loader version
0211 0227/1 2.02+(3) ext_loader_type Extended boot loader ID
0212 0228/4 2.02+ cmd_line_ptr 32-bit pointer to the kernel command line
0213 022C/4 2.03+ initrd_addr_max Highest legal initrd address
0214 0230/4 2.05+ kernel_alignment Physical addr alignment required for kernel
0215 0234/1 2.05+ relocatable_kernel Whether kernel is relocatable or not
0216 0235/1 2.10+ min_alignment Minimum alignment, as a power of two
0217 0236/2 2.12+ xloadflags Boot protocol option flags
0218 0238/4 2.06+ cmdline_size Maximum size of the kernel command line
0219 023C/4 2.07+ hardware_subarch Hardware subarchitecture
0220 0240/8 2.07+ hardware_subarch_data Subarchitecture-specific data
0221 0248/4 2.08+ payload_offset Offset of kernel payload
0222 024C/4 2.08+ payload_length Length of kernel payload
0223 0250/8 2.09+ setup_data 64-bit physical pointer to linked list
0224 of struct setup_data
0225 0258/8 2.10+ pref_address Preferred loading address
0226 0260/4 2.10+ init_size Linear memory required during initialization
0227 0264/4 2.11+ handover_offset Offset of handover entry point
0228 0268/4 2.15+ kernel_info_offset Offset of the kernel_info
0229 =========== ======== ===================== ============================================
0230
0231 .. note::
0232 (1) For backwards compatibility, if the setup_sects field contains 0, the
0233 real value is 4.
0234
0235 (2) For boot protocol prior to 2.04, the upper two bytes of the syssize
0236 field are unusable, which means the size of a bzImage kernel
0237 cannot be determined.
0238
0239 (3) Ignored, but safe to set, for boot protocols 2.02-2.09.
0240
0241 If the "HdrS" (0x53726448) magic number is not found at offset 0x202,
0242 the boot protocol version is "old". Loading an old kernel, the
0243 following parameters should be assumed::
0244
0245 Image type = zImage
0246 initrd not supported
0247 Real-mode kernel must be located at 0x90000.
0248
0249 Otherwise, the "version" field contains the protocol version,
0250 e.g. protocol version 2.01 will contain 0x0201 in this field. When
0251 setting fields in the header, you must make sure only to set fields
0252 supported by the protocol version in use.
0253
0254
0255 Details of Header Fields
0256 ========================
0257
0258 For each field, some are information from the kernel to the bootloader
0259 ("read"), some are expected to be filled out by the bootloader
0260 ("write"), and some are expected to be read and modified by the
0261 bootloader ("modify").
0262
0263 All general purpose boot loaders should write the fields marked
0264 (obligatory). Boot loaders who want to load the kernel at a
0265 nonstandard address should fill in the fields marked (reloc); other
0266 boot loaders can ignore those fields.
0267
0268 The byte order of all fields is littleendian (this is x86, after all.)
0269
0270 ============ ===========
0271 Field name: setup_sects
0272 Type: read
0273 Offset/size: 0x1f1/1
0274 Protocol: ALL
0275 ============ ===========
0276
0277 The size of the setup code in 512-byte sectors. If this field is
0278 0, the real value is 4. The real-mode code consists of the boot
0279 sector (always one 512-byte sector) plus the setup code.
0280
0281 ============ =================
0282 Field name: root_flags
0283 Type: modify (optional)
0284 Offset/size: 0x1f2/2
0285 Protocol: ALL
0286 ============ =================
0287
0288 If this field is nonzero, the root defaults to readonly. The use of
0289 this field is deprecated; use the "ro" or "rw" options on the
0290 command line instead.
0291
0292 ============ ===============================================
0293 Field name: syssize
0294 Type: read
0295 Offset/size: 0x1f4/4 (protocol 2.04+) 0x1f4/2 (protocol ALL)
0296 Protocol: 2.04+
0297 ============ ===============================================
0298
0299 The size of the protected-mode code in units of 16-byte paragraphs.
0300 For protocol versions older than 2.04 this field is only two bytes
0301 wide, and therefore cannot be trusted for the size of a kernel if
0302 the LOAD_HIGH flag is set.
0303
0304 ============ ===============
0305 Field name: ram_size
0306 Type: kernel internal
0307 Offset/size: 0x1f8/2
0308 Protocol: ALL
0309 ============ ===============
0310
0311 This field is obsolete.
0312
0313 ============ ===================
0314 Field name: vid_mode
0315 Type: modify (obligatory)
0316 Offset/size: 0x1fa/2
0317 ============ ===================
0318
0319 Please see the section on SPECIAL COMMAND LINE OPTIONS.
0320
0321 ============ =================
0322 Field name: root_dev
0323 Type: modify (optional)
0324 Offset/size: 0x1fc/2
0325 Protocol: ALL
0326 ============ =================
0327
0328 The default root device device number. The use of this field is
0329 deprecated, use the "root=" option on the command line instead.
0330
0331 ============ =========
0332 Field name: boot_flag
0333 Type: read
0334 Offset/size: 0x1fe/2
0335 Protocol: ALL
0336 ============ =========
0337
0338 Contains 0xAA55. This is the closest thing old Linux kernels have
0339 to a magic number.
0340
0341 ============ =======
0342 Field name: jump
0343 Type: read
0344 Offset/size: 0x200/2
0345 Protocol: 2.00+
0346 ============ =======
0347
0348 Contains an x86 jump instruction, 0xEB followed by a signed offset
0349 relative to byte 0x202. This can be used to determine the size of
0350 the header.
0351
0352 ============ =======
0353 Field name: header
0354 Type: read
0355 Offset/size: 0x202/4
0356 Protocol: 2.00+
0357 ============ =======
0358
0359 Contains the magic number "HdrS" (0x53726448).
0360
0361 ============ =======
0362 Field name: version
0363 Type: read
0364 Offset/size: 0x206/2
0365 Protocol: 2.00+
0366 ============ =======
0367
0368 Contains the boot protocol version, in (major << 8)+minor format,
0369 e.g. 0x0204 for version 2.04, and 0x0a11 for a hypothetical version
0370 10.17.
0371
0372 ============ =================
0373 Field name: realmode_swtch
0374 Type: modify (optional)
0375 Offset/size: 0x208/4
0376 Protocol: 2.00+
0377 ============ =================
0378
0379 Boot loader hook (see ADVANCED BOOT LOADER HOOKS below.)
0380
0381 ============ =============
0382 Field name: start_sys_seg
0383 Type: read
0384 Offset/size: 0x20c/2
0385 Protocol: 2.00+
0386 ============ =============
0387
0388 The load low segment (0x1000). Obsolete.
0389
0390 ============ ==============
0391 Field name: kernel_version
0392 Type: read
0393 Offset/size: 0x20e/2
0394 Protocol: 2.00+
0395 ============ ==============
0396
0397 If set to a nonzero value, contains a pointer to a NUL-terminated
0398 human-readable kernel version number string, less 0x200. This can
0399 be used to display the kernel version to the user. This value
0400 should be less than (0x200*setup_sects).
0401
0402 For example, if this value is set to 0x1c00, the kernel version
0403 number string can be found at offset 0x1e00 in the kernel file.
0404 This is a valid value if and only if the "setup_sects" field
0405 contains the value 15 or higher, as::
0406
0407 0x1c00 < 15*0x200 (= 0x1e00) but
0408 0x1c00 >= 14*0x200 (= 0x1c00)
0409
0410 0x1c00 >> 9 = 14, So the minimum value for setup_secs is 15.
0411
0412 ============ ==================
0413 Field name: type_of_loader
0414 Type: write (obligatory)
0415 Offset/size: 0x210/1
0416 Protocol: 2.00+
0417 ============ ==================
0418
0419 If your boot loader has an assigned id (see table below), enter
0420 0xTV here, where T is an identifier for the boot loader and V is
0421 a version number. Otherwise, enter 0xFF here.
0422
0423 For boot loader IDs above T = 0xD, write T = 0xE to this field and
0424 write the extended ID minus 0x10 to the ext_loader_type field.
0425 Similarly, the ext_loader_ver field can be used to provide more than
0426 four bits for the bootloader version.
0427
0428 For example, for T = 0x15, V = 0x234, write::
0429
0430 type_of_loader <- 0xE4
0431 ext_loader_type <- 0x05
0432 ext_loader_ver <- 0x23
0433
0434 Assigned boot loader ids (hexadecimal):
0435
0436 == =======================================
0437 0 LILO
0438 (0x00 reserved for pre-2.00 bootloader)
0439 1 Loadlin
0440 2 bootsect-loader
0441 (0x20, all other values reserved)
0442 3 Syslinux
0443 4 Etherboot/gPXE/iPXE
0444 5 ELILO
0445 7 GRUB
0446 8 U-Boot
0447 9 Xen
0448 A Gujin
0449 B Qemu
0450 C Arcturus Networks uCbootloader
0451 D kexec-tools
0452 E Extended (see ext_loader_type)
0453 F Special (0xFF = undefined)
0454 10 Reserved
0455 11 Minimal Linux Bootloader
0456 <http://sebastian-plotz.blogspot.de>
0457 12 OVMF UEFI virtualization stack
0458 == =======================================
0459
0460 Please contact <hpa@zytor.com> if you need a bootloader ID value assigned.
0461
0462 ============ ===================
0463 Field name: loadflags
0464 Type: modify (obligatory)
0465 Offset/size: 0x211/1
0466 Protocol: 2.00+
0467 ============ ===================
0468
0469 This field is a bitmask.
0470
0471 Bit 0 (read): LOADED_HIGH
0472
0473 - If 0, the protected-mode code is loaded at 0x10000.
0474 - If 1, the protected-mode code is loaded at 0x100000.
0475
0476 Bit 1 (kernel internal): KASLR_FLAG
0477
0478 - Used internally by the compressed kernel to communicate
0479 KASLR status to kernel proper.
0480
0481 - If 1, KASLR enabled.
0482 - If 0, KASLR disabled.
0483
0484 Bit 5 (write): QUIET_FLAG
0485
0486 - If 0, print early messages.
0487 - If 1, suppress early messages.
0488
0489 This requests to the kernel (decompressor and early
0490 kernel) to not write early messages that require
0491 accessing the display hardware directly.
0492
0493 Bit 6 (obsolete): KEEP_SEGMENTS
0494
0495 Protocol: 2.07+
0496
0497 - This flag is obsolete.
0498
0499 Bit 7 (write): CAN_USE_HEAP
0500
0501 Set this bit to 1 to indicate that the value entered in the
0502 heap_end_ptr is valid. If this field is clear, some setup code
0503 functionality will be disabled.
0504
0505
0506 ============ ===================
0507 Field name: setup_move_size
0508 Type: modify (obligatory)
0509 Offset/size: 0x212/2
0510 Protocol: 2.00-2.01
0511 ============ ===================
0512
0513 When using protocol 2.00 or 2.01, if the real mode kernel is not
0514 loaded at 0x90000, it gets moved there later in the loading
0515 sequence. Fill in this field if you want additional data (such as
0516 the kernel command line) moved in addition to the real-mode kernel
0517 itself.
0518
0519 The unit is bytes starting with the beginning of the boot sector.
0520
0521 This field is can be ignored when the protocol is 2.02 or higher, or
0522 if the real-mode code is loaded at 0x90000.
0523
0524 ============ ========================
0525 Field name: code32_start
0526 Type: modify (optional, reloc)
0527 Offset/size: 0x214/4
0528 Protocol: 2.00+
0529 ============ ========================
0530
0531 The address to jump to in protected mode. This defaults to the load
0532 address of the kernel, and can be used by the boot loader to
0533 determine the proper load address.
0534
0535 This field can be modified for two purposes:
0536
0537 1. as a boot loader hook (see Advanced Boot Loader Hooks below.)
0538
0539 2. if a bootloader which does not install a hook loads a
0540 relocatable kernel at a nonstandard address it will have to modify
0541 this field to point to the load address.
0542
0543 ============ ==================
0544 Field name: ramdisk_image
0545 Type: write (obligatory)
0546 Offset/size: 0x218/4
0547 Protocol: 2.00+
0548 ============ ==================
0549
0550 The 32-bit linear address of the initial ramdisk or ramfs. Leave at
0551 zero if there is no initial ramdisk/ramfs.
0552
0553 ============ ==================
0554 Field name: ramdisk_size
0555 Type: write (obligatory)
0556 Offset/size: 0x21c/4
0557 Protocol: 2.00+
0558 ============ ==================
0559
0560 Size of the initial ramdisk or ramfs. Leave at zero if there is no
0561 initial ramdisk/ramfs.
0562
0563 ============ ===============
0564 Field name: bootsect_kludge
0565 Type: kernel internal
0566 Offset/size: 0x220/4
0567 Protocol: 2.00+
0568 ============ ===============
0569
0570 This field is obsolete.
0571
0572 ============ ==================
0573 Field name: heap_end_ptr
0574 Type: write (obligatory)
0575 Offset/size: 0x224/2
0576 Protocol: 2.01+
0577 ============ ==================
0578
0579 Set this field to the offset (from the beginning of the real-mode
0580 code) of the end of the setup stack/heap, minus 0x0200.
0581
0582 ============ ================
0583 Field name: ext_loader_ver
0584 Type: write (optional)
0585 Offset/size: 0x226/1
0586 Protocol: 2.02+
0587 ============ ================
0588
0589 This field is used as an extension of the version number in the
0590 type_of_loader field. The total version number is considered to be
0591 (type_of_loader & 0x0f) + (ext_loader_ver << 4).
0592
0593 The use of this field is boot loader specific. If not written, it
0594 is zero.
0595
0596 Kernels prior to 2.6.31 did not recognize this field, but it is safe
0597 to write for protocol version 2.02 or higher.
0598
0599 ============ =====================================================
0600 Field name: ext_loader_type
0601 Type: write (obligatory if (type_of_loader & 0xf0) == 0xe0)
0602 Offset/size: 0x227/1
0603 Protocol: 2.02+
0604 ============ =====================================================
0605
0606 This field is used as an extension of the type number in
0607 type_of_loader field. If the type in type_of_loader is 0xE, then
0608 the actual type is (ext_loader_type + 0x10).
0609
0610 This field is ignored if the type in type_of_loader is not 0xE.
0611
0612 Kernels prior to 2.6.31 did not recognize this field, but it is safe
0613 to write for protocol version 2.02 or higher.
0614
0615 ============ ==================
0616 Field name: cmd_line_ptr
0617 Type: write (obligatory)
0618 Offset/size: 0x228/4
0619 Protocol: 2.02+
0620 ============ ==================
0621
0622 Set this field to the linear address of the kernel command line.
0623 The kernel command line can be located anywhere between the end of
0624 the setup heap and 0xA0000; it does not have to be located in the
0625 same 64K segment as the real-mode code itself.
0626
0627 Fill in this field even if your boot loader does not support a
0628 command line, in which case you can point this to an empty string
0629 (or better yet, to the string "auto".) If this field is left at
0630 zero, the kernel will assume that your boot loader does not support
0631 the 2.02+ protocol.
0632
0633 ============ ===============
0634 Field name: initrd_addr_max
0635 Type: read
0636 Offset/size: 0x22c/4
0637 Protocol: 2.03+
0638 ============ ===============
0639
0640 The maximum address that may be occupied by the initial
0641 ramdisk/ramfs contents. For boot protocols 2.02 or earlier, this
0642 field is not present, and the maximum address is 0x37FFFFFF. (This
0643 address is defined as the address of the highest safe byte, so if
0644 your ramdisk is exactly 131072 bytes long and this field is
0645 0x37FFFFFF, you can start your ramdisk at 0x37FE0000.)
0646
0647 ============ ============================
0648 Field name: kernel_alignment
0649 Type: read/modify (reloc)
0650 Offset/size: 0x230/4
0651 Protocol: 2.05+ (read), 2.10+ (modify)
0652 ============ ============================
0653
0654 Alignment unit required by the kernel (if relocatable_kernel is
0655 true.) A relocatable kernel that is loaded at an alignment
0656 incompatible with the value in this field will be realigned during
0657 kernel initialization.
0658
0659 Starting with protocol version 2.10, this reflects the kernel
0660 alignment preferred for optimal performance; it is possible for the
0661 loader to modify this field to permit a lesser alignment. See the
0662 min_alignment and pref_address field below.
0663
0664 ============ ==================
0665 Field name: relocatable_kernel
0666 Type: read (reloc)
0667 Offset/size: 0x234/1
0668 Protocol: 2.05+
0669 ============ ==================
0670
0671 If this field is nonzero, the protected-mode part of the kernel can
0672 be loaded at any address that satisfies the kernel_alignment field.
0673 After loading, the boot loader must set the code32_start field to
0674 point to the loaded code, or to a boot loader hook.
0675
0676 ============ =============
0677 Field name: min_alignment
0678 Type: read (reloc)
0679 Offset/size: 0x235/1
0680 Protocol: 2.10+
0681 ============ =============
0682
0683 This field, if nonzero, indicates as a power of two the minimum
0684 alignment required, as opposed to preferred, by the kernel to boot.
0685 If a boot loader makes use of this field, it should update the
0686 kernel_alignment field with the alignment unit desired; typically::
0687
0688 kernel_alignment = 1 << min_alignment
0689
0690 There may be a considerable performance cost with an excessively
0691 misaligned kernel. Therefore, a loader should typically try each
0692 power-of-two alignment from kernel_alignment down to this alignment.
0693
0694 ============ ==========
0695 Field name: xloadflags
0696 Type: read
0697 Offset/size: 0x236/2
0698 Protocol: 2.12+
0699 ============ ==========
0700
0701 This field is a bitmask.
0702
0703 Bit 0 (read): XLF_KERNEL_64
0704
0705 - If 1, this kernel has the legacy 64-bit entry point at 0x200.
0706
0707 Bit 1 (read): XLF_CAN_BE_LOADED_ABOVE_4G
0708
0709 - If 1, kernel/boot_params/cmdline/ramdisk can be above 4G.
0710
0711 Bit 2 (read): XLF_EFI_HANDOVER_32
0712
0713 - If 1, the kernel supports the 32-bit EFI handoff entry point
0714 given at handover_offset.
0715
0716 Bit 3 (read): XLF_EFI_HANDOVER_64
0717
0718 - If 1, the kernel supports the 64-bit EFI handoff entry point
0719 given at handover_offset + 0x200.
0720
0721 Bit 4 (read): XLF_EFI_KEXEC
0722
0723 - If 1, the kernel supports kexec EFI boot with EFI runtime support.
0724
0725
0726 ============ ============
0727 Field name: cmdline_size
0728 Type: read
0729 Offset/size: 0x238/4
0730 Protocol: 2.06+
0731 ============ ============
0732
0733 The maximum size of the command line without the terminating
0734 zero. This means that the command line can contain at most
0735 cmdline_size characters. With protocol version 2.05 and earlier, the
0736 maximum size was 255.
0737
0738 ============ ====================================
0739 Field name: hardware_subarch
0740 Type: write (optional, defaults to x86/PC)
0741 Offset/size: 0x23c/4
0742 Protocol: 2.07+
0743 ============ ====================================
0744
0745 In a paravirtualized environment the hardware low level architectural
0746 pieces such as interrupt handling, page table handling, and
0747 accessing process control registers needs to be done differently.
0748
0749 This field allows the bootloader to inform the kernel we are in one
0750 one of those environments.
0751
0752 ========== ==============================
0753 0x00000000 The default x86/PC environment
0754 0x00000001 lguest
0755 0x00000002 Xen
0756 0x00000003 Moorestown MID
0757 0x00000004 CE4100 TV Platform
0758 ========== ==============================
0759
0760 ============ =========================
0761 Field name: hardware_subarch_data
0762 Type: write (subarch-dependent)
0763 Offset/size: 0x240/8
0764 Protocol: 2.07+
0765 ============ =========================
0766
0767 A pointer to data that is specific to hardware subarch
0768 This field is currently unused for the default x86/PC environment,
0769 do not modify.
0770
0771 ============ ==============
0772 Field name: payload_offset
0773 Type: read
0774 Offset/size: 0x248/4
0775 Protocol: 2.08+
0776 ============ ==============
0777
0778 If non-zero then this field contains the offset from the beginning
0779 of the protected-mode code to the payload.
0780
0781 The payload may be compressed. The format of both the compressed and
0782 uncompressed data should be determined using the standard magic
0783 numbers. The currently supported compression formats are gzip
0784 (magic numbers 1F 8B or 1F 9E), bzip2 (magic number 42 5A), LZMA
0785 (magic number 5D 00), XZ (magic number FD 37), LZ4 (magic number
0786 02 21) and ZSTD (magic number 28 B5). The uncompressed payload is
0787 currently always ELF (magic number 7F 45 4C 46).
0788
0789 ============ ==============
0790 Field name: payload_length
0791 Type: read
0792 Offset/size: 0x24c/4
0793 Protocol: 2.08+
0794 ============ ==============
0795
0796 The length of the payload.
0797
0798 ============ ===============
0799 Field name: setup_data
0800 Type: write (special)
0801 Offset/size: 0x250/8
0802 Protocol: 2.09+
0803 ============ ===============
0804
0805 The 64-bit physical pointer to NULL terminated single linked list of
0806 struct setup_data. This is used to define a more extensible boot
0807 parameters passing mechanism. The definition of struct setup_data is
0808 as follow::
0809
0810 struct setup_data {
0811 u64 next;
0812 u32 type;
0813 u32 len;
0814 u8 data[0];
0815 };
0816
0817 Where, the next is a 64-bit physical pointer to the next node of
0818 linked list, the next field of the last node is 0; the type is used
0819 to identify the contents of data; the len is the length of data
0820 field; the data holds the real payload.
0821
0822 This list may be modified at a number of points during the bootup
0823 process. Therefore, when modifying this list one should always make
0824 sure to consider the case where the linked list already contains
0825 entries.
0826
0827 The setup_data is a bit awkward to use for extremely large data objects,
0828 both because the setup_data header has to be adjacent to the data object
0829 and because it has a 32-bit length field. However, it is important that
0830 intermediate stages of the boot process have a way to identify which
0831 chunks of memory are occupied by kernel data.
0832
0833 Thus setup_indirect struct and SETUP_INDIRECT type were introduced in
0834 protocol 2.15::
0835
0836 struct setup_indirect {
0837 __u32 type;
0838 __u32 reserved; /* Reserved, must be set to zero. */
0839 __u64 len;
0840 __u64 addr;
0841 };
0842
0843 The type member is a SETUP_INDIRECT | SETUP_* type. However, it cannot be
0844 SETUP_INDIRECT itself since making the setup_indirect a tree structure
0845 could require a lot of stack space in something that needs to parse it
0846 and stack space can be limited in boot contexts.
0847
0848 Let's give an example how to point to SETUP_E820_EXT data using setup_indirect.
0849 In this case setup_data and setup_indirect will look like this::
0850
0851 struct setup_data {
0852 __u64 next = 0 or <addr_of_next_setup_data_struct>;
0853 __u32 type = SETUP_INDIRECT;
0854 __u32 len = sizeof(setup_indirect);
0855 __u8 data[sizeof(setup_indirect)] = struct setup_indirect {
0856 __u32 type = SETUP_INDIRECT | SETUP_E820_EXT;
0857 __u32 reserved = 0;
0858 __u64 len = <len_of_SETUP_E820_EXT_data>;
0859 __u64 addr = <addr_of_SETUP_E820_EXT_data>;
0860 }
0861 }
0862
0863 .. note::
0864 SETUP_INDIRECT | SETUP_NONE objects cannot be properly distinguished
0865 from SETUP_INDIRECT itself. So, this kind of objects cannot be provided
0866 by the bootloaders.
0867
0868 ============ ============
0869 Field name: pref_address
0870 Type: read (reloc)
0871 Offset/size: 0x258/8
0872 Protocol: 2.10+
0873 ============ ============
0874
0875 This field, if nonzero, represents a preferred load address for the
0876 kernel. A relocating bootloader should attempt to load at this
0877 address if possible.
0878
0879 A non-relocatable kernel will unconditionally move itself and to run
0880 at this address.
0881
0882 ============ =======
0883 Field name: init_size
0884 Type: read
0885 Offset/size: 0x260/4
0886 ============ =======
0887
0888 This field indicates the amount of linear contiguous memory starting
0889 at the kernel runtime start address that the kernel needs before it
0890 is capable of examining its memory map. This is not the same thing
0891 as the total amount of memory the kernel needs to boot, but it can
0892 be used by a relocating boot loader to help select a safe load
0893 address for the kernel.
0894
0895 The kernel runtime start address is determined by the following algorithm::
0896
0897 if (relocatable_kernel)
0898 runtime_start = align_up(load_address, kernel_alignment)
0899 else
0900 runtime_start = pref_address
0901
0902 ============ ===============
0903 Field name: handover_offset
0904 Type: read
0905 Offset/size: 0x264/4
0906 ============ ===============
0907
0908 This field is the offset from the beginning of the kernel image to
0909 the EFI handover protocol entry point. Boot loaders using the EFI
0910 handover protocol to boot the kernel should jump to this offset.
0911
0912 See EFI HANDOVER PROTOCOL below for more details.
0913
0914 ============ ==================
0915 Field name: kernel_info_offset
0916 Type: read
0917 Offset/size: 0x268/4
0918 Protocol: 2.15+
0919 ============ ==================
0920
0921 This field is the offset from the beginning of the kernel image to the
0922 kernel_info. The kernel_info structure is embedded in the Linux image
0923 in the uncompressed protected mode region.
0924
0925
0926 The kernel_info
0927 ===============
0928
0929 The relationships between the headers are analogous to the various data
0930 sections:
0931
0932 setup_header = .data
0933 boot_params/setup_data = .bss
0934
0935 What is missing from the above list? That's right:
0936
0937 kernel_info = .rodata
0938
0939 We have been (ab)using .data for things that could go into .rodata or .bss for
0940 a long time, for lack of alternatives and -- especially early on -- inertia.
0941 Also, the BIOS stub is responsible for creating boot_params, so it isn't
0942 available to a BIOS-based loader (setup_data is, though).
0943
0944 setup_header is permanently limited to 144 bytes due to the reach of the
0945 2-byte jump field, which doubles as a length field for the structure, combined
0946 with the size of the "hole" in struct boot_params that a protected-mode loader
0947 or the BIOS stub has to copy it into. It is currently 119 bytes long, which
0948 leaves us with 25 very precious bytes. This isn't something that can be fixed
0949 without revising the boot protocol entirely, breaking backwards compatibility.
0950
0951 boot_params proper is limited to 4096 bytes, but can be arbitrarily extended
0952 by adding setup_data entries. It cannot be used to communicate properties of
0953 the kernel image, because it is .bss and has no image-provided content.
0954
0955 kernel_info solves this by providing an extensible place for information about
0956 the kernel image. It is readonly, because the kernel cannot rely on a
0957 bootloader copying its contents anywhere, but that is OK; if it becomes
0958 necessary it can still contain data items that an enabled bootloader would be
0959 expected to copy into a setup_data chunk.
0960
0961 All kernel_info data should be part of this structure. Fixed size data have to
0962 be put before kernel_info_var_len_data label. Variable size data have to be put
0963 after kernel_info_var_len_data label. Each chunk of variable size data has to
0964 be prefixed with header/magic and its size, e.g.::
0965
0966 kernel_info:
0967 .ascii "LToP" /* Header, Linux top (structure). */
0968 .long kernel_info_var_len_data - kernel_info
0969 .long kernel_info_end - kernel_info
0970 .long 0x01234567 /* Some fixed size data for the bootloaders. */
0971 kernel_info_var_len_data:
0972 example_struct: /* Some variable size data for the bootloaders. */
0973 .ascii "0123" /* Header/Magic. */
0974 .long example_struct_end - example_struct
0975 .ascii "Struct"
0976 .long 0x89012345
0977 example_struct_end:
0978 example_strings: /* Some variable size data for the bootloaders. */
0979 .ascii "ABCD" /* Header/Magic. */
0980 .long example_strings_end - example_strings
0981 .asciz "String_0"
0982 .asciz "String_1"
0983 example_strings_end:
0984 kernel_info_end:
0985
0986 This way the kernel_info is self-contained blob.
0987
0988 .. note::
0989 Each variable size data header/magic can be any 4-character string,
0990 without \0 at the end of the string, which does not collide with
0991 existing variable length data headers/magics.
0992
0993
0994 Details of the kernel_info Fields
0995 =================================
0996
0997 ============ ========
0998 Field name: header
0999 Offset/size: 0x0000/4
1000 ============ ========
1001
1002 Contains the magic number "LToP" (0x506f544c).
1003
1004 ============ ========
1005 Field name: size
1006 Offset/size: 0x0004/4
1007 ============ ========
1008
1009 This field contains the size of the kernel_info including kernel_info.header.
1010 It does not count kernel_info.kernel_info_var_len_data size. This field should be
1011 used by the bootloaders to detect supported fixed size fields in the kernel_info
1012 and beginning of kernel_info.kernel_info_var_len_data.
1013
1014 ============ ========
1015 Field name: size_total
1016 Offset/size: 0x0008/4
1017 ============ ========
1018
1019 This field contains the size of the kernel_info including kernel_info.header
1020 and kernel_info.kernel_info_var_len_data.
1021
1022 ============ ==============
1023 Field name: setup_type_max
1024 Offset/size: 0x000c/4
1025 ============ ==============
1026
1027 This field contains maximal allowed type for setup_data and setup_indirect structs.
1028
1029
1030 The Image Checksum
1031 ==================
1032
1033 From boot protocol version 2.08 onwards the CRC-32 is calculated over
1034 the entire file using the characteristic polynomial 0x04C11DB7 and an
1035 initial remainder of 0xffffffff. The checksum is appended to the
1036 file; therefore the CRC of the file up to the limit specified in the
1037 syssize field of the header is always 0.
1038
1039
1040 The Kernel Command Line
1041 =======================
1042
1043 The kernel command line has become an important way for the boot
1044 loader to communicate with the kernel. Some of its options are also
1045 relevant to the boot loader itself, see "special command line options"
1046 below.
1047
1048 The kernel command line is a null-terminated string. The maximum
1049 length can be retrieved from the field cmdline_size. Before protocol
1050 version 2.06, the maximum was 255 characters. A string that is too
1051 long will be automatically truncated by the kernel.
1052
1053 If the boot protocol version is 2.02 or later, the address of the
1054 kernel command line is given by the header field cmd_line_ptr (see
1055 above.) This address can be anywhere between the end of the setup
1056 heap and 0xA0000.
1057
1058 If the protocol version is *not* 2.02 or higher, the kernel
1059 command line is entered using the following protocol:
1060
1061 - At offset 0x0020 (word), "cmd_line_magic", enter the magic
1062 number 0xA33F.
1063
1064 - At offset 0x0022 (word), "cmd_line_offset", enter the offset
1065 of the kernel command line (relative to the start of the
1066 real-mode kernel).
1067
1068 - The kernel command line *must* be within the memory region
1069 covered by setup_move_size, so you may need to adjust this
1070 field.
1071
1072
1073 Memory Layout of The Real-Mode Code
1074 ===================================
1075
1076 The real-mode code requires a stack/heap to be set up, as well as
1077 memory allocated for the kernel command line. This needs to be done
1078 in the real-mode accessible memory in bottom megabyte.
1079
1080 It should be noted that modern machines often have a sizable Extended
1081 BIOS Data Area (EBDA). As a result, it is advisable to use as little
1082 of the low megabyte as possible.
1083
1084 Unfortunately, under the following circumstances the 0x90000 memory
1085 segment has to be used:
1086
1087 - When loading a zImage kernel ((loadflags & 0x01) == 0).
1088 - When loading a 2.01 or earlier boot protocol kernel.
1089
1090 .. note::
1091 For the 2.00 and 2.01 boot protocols, the real-mode code
1092 can be loaded at another address, but it is internally
1093 relocated to 0x90000. For the "old" protocol, the
1094 real-mode code must be loaded at 0x90000.
1095
1096 When loading at 0x90000, avoid using memory above 0x9a000.
1097
1098 For boot protocol 2.02 or higher, the command line does not have to be
1099 located in the same 64K segment as the real-mode setup code; it is
1100 thus permitted to give the stack/heap the full 64K segment and locate
1101 the command line above it.
1102
1103 The kernel command line should not be located below the real-mode
1104 code, nor should it be located in high memory.
1105
1106
1107 Sample Boot Configuartion
1108 =========================
1109
1110 As a sample configuration, assume the following layout of the real
1111 mode segment.
1112
1113 When loading below 0x90000, use the entire segment:
1114
1115 ============= ===================
1116 0x0000-0x7fff Real mode kernel
1117 0x8000-0xdfff Stack and heap
1118 0xe000-0xffff Kernel command line
1119 ============= ===================
1120
1121 When loading at 0x90000 OR the protocol version is 2.01 or earlier:
1122
1123 ============= ===================
1124 0x0000-0x7fff Real mode kernel
1125 0x8000-0x97ff Stack and heap
1126 0x9800-0x9fff Kernel command line
1127 ============= ===================
1128
1129 Such a boot loader should enter the following fields in the header::
1130
1131 unsigned long base_ptr; /* base address for real-mode segment */
1132
1133 if ( setup_sects == 0 ) {
1134 setup_sects = 4;
1135 }
1136
1137 if ( protocol >= 0x0200 ) {
1138 type_of_loader = <type code>;
1139 if ( loading_initrd ) {
1140 ramdisk_image = <initrd_address>;
1141 ramdisk_size = <initrd_size>;
1142 }
1143
1144 if ( protocol >= 0x0202 && loadflags & 0x01 )
1145 heap_end = 0xe000;
1146 else
1147 heap_end = 0x9800;
1148
1149 if ( protocol >= 0x0201 ) {
1150 heap_end_ptr = heap_end - 0x200;
1151 loadflags |= 0x80; /* CAN_USE_HEAP */
1152 }
1153
1154 if ( protocol >= 0x0202 ) {
1155 cmd_line_ptr = base_ptr + heap_end;
1156 strcpy(cmd_line_ptr, cmdline);
1157 } else {
1158 cmd_line_magic = 0xA33F;
1159 cmd_line_offset = heap_end;
1160 setup_move_size = heap_end + strlen(cmdline)+1;
1161 strcpy(base_ptr+cmd_line_offset, cmdline);
1162 }
1163 } else {
1164 /* Very old kernel */
1165
1166 heap_end = 0x9800;
1167
1168 cmd_line_magic = 0xA33F;
1169 cmd_line_offset = heap_end;
1170
1171 /* A very old kernel MUST have its real-mode code
1172 loaded at 0x90000 */
1173
1174 if ( base_ptr != 0x90000 ) {
1175 /* Copy the real-mode kernel */
1176 memcpy(0x90000, base_ptr, (setup_sects+1)*512);
1177 base_ptr = 0x90000; /* Relocated */
1178 }
1179
1180 strcpy(0x90000+cmd_line_offset, cmdline);
1181
1182 /* It is recommended to clear memory up to the 32K mark */
1183 memset(0x90000 + (setup_sects+1)*512, 0,
1184 (64-(setup_sects+1))*512);
1185 }
1186
1187
1188 Loading The Rest of The Kernel
1189 ==============================
1190
1191 The 32-bit (non-real-mode) kernel starts at offset (setup_sects+1)*512
1192 in the kernel file (again, if setup_sects == 0 the real value is 4.)
1193 It should be loaded at address 0x10000 for Image/zImage kernels and
1194 0x100000 for bzImage kernels.
1195
1196 The kernel is a bzImage kernel if the protocol >= 2.00 and the 0x01
1197 bit (LOAD_HIGH) in the loadflags field is set::
1198
1199 is_bzImage = (protocol >= 0x0200) && (loadflags & 0x01);
1200 load_address = is_bzImage ? 0x100000 : 0x10000;
1201
1202 Note that Image/zImage kernels can be up to 512K in size, and thus use
1203 the entire 0x10000-0x90000 range of memory. This means it is pretty
1204 much a requirement for these kernels to load the real-mode part at
1205 0x90000. bzImage kernels allow much more flexibility.
1206
1207 Special Command Line Options
1208 ============================
1209
1210 If the command line provided by the boot loader is entered by the
1211 user, the user may expect the following command line options to work.
1212 They should normally not be deleted from the kernel command line even
1213 though not all of them are actually meaningful to the kernel. Boot
1214 loader authors who need additional command line options for the boot
1215 loader itself should get them registered in
1216 Documentation/admin-guide/kernel-parameters.rst to make sure they will not
1217 conflict with actual kernel options now or in the future.
1218
1219 vga=<mode>
1220 <mode> here is either an integer (in C notation, either
1221 decimal, octal, or hexadecimal) or one of the strings
1222 "normal" (meaning 0xFFFF), "ext" (meaning 0xFFFE) or "ask"
1223 (meaning 0xFFFD). This value should be entered into the
1224 vid_mode field, as it is used by the kernel before the command
1225 line is parsed.
1226
1227 mem=<size>
1228 <size> is an integer in C notation optionally followed by
1229 (case insensitive) K, M, G, T, P or E (meaning << 10, << 20,
1230 << 30, << 40, << 50 or << 60). This specifies the end of
1231 memory to the kernel. This affects the possible placement of
1232 an initrd, since an initrd should be placed near end of
1233 memory. Note that this is an option to *both* the kernel and
1234 the bootloader!
1235
1236 initrd=<file>
1237 An initrd should be loaded. The meaning of <file> is
1238 obviously bootloader-dependent, and some boot loaders
1239 (e.g. LILO) do not have such a command.
1240
1241 In addition, some boot loaders add the following options to the
1242 user-specified command line:
1243
1244 BOOT_IMAGE=<file>
1245 The boot image which was loaded. Again, the meaning of <file>
1246 is obviously bootloader-dependent.
1247
1248 auto
1249 The kernel was booted without explicit user intervention.
1250
1251 If these options are added by the boot loader, it is highly
1252 recommended that they are located *first*, before the user-specified
1253 or configuration-specified command line. Otherwise, "init=/bin/sh"
1254 gets confused by the "auto" option.
1255
1256
1257 Running the Kernel
1258 ==================
1259
1260 The kernel is started by jumping to the kernel entry point, which is
1261 located at *segment* offset 0x20 from the start of the real mode
1262 kernel. This means that if you loaded your real-mode kernel code at
1263 0x90000, the kernel entry point is 9020:0000.
1264
1265 At entry, ds = es = ss should point to the start of the real-mode
1266 kernel code (0x9000 if the code is loaded at 0x90000), sp should be
1267 set up properly, normally pointing to the top of the heap, and
1268 interrupts should be disabled. Furthermore, to guard against bugs in
1269 the kernel, it is recommended that the boot loader sets fs = gs = ds =
1270 es = ss.
1271
1272 In our example from above, we would do::
1273
1274 /* Note: in the case of the "old" kernel protocol, base_ptr must
1275 be == 0x90000 at this point; see the previous sample code */
1276
1277 seg = base_ptr >> 4;
1278
1279 cli(); /* Enter with interrupts disabled! */
1280
1281 /* Set up the real-mode kernel stack */
1282 _SS = seg;
1283 _SP = heap_end;
1284
1285 _DS = _ES = _FS = _GS = seg;
1286 jmp_far(seg+0x20, 0); /* Run the kernel */
1287
1288 If your boot sector accesses a floppy drive, it is recommended to
1289 switch off the floppy motor before running the kernel, since the
1290 kernel boot leaves interrupts off and thus the motor will not be
1291 switched off, especially if the loaded kernel has the floppy driver as
1292 a demand-loaded module!
1293
1294
1295 Advanced Boot Loader Hooks
1296 ==========================
1297
1298 If the boot loader runs in a particularly hostile environment (such as
1299 LOADLIN, which runs under DOS) it may be impossible to follow the
1300 standard memory location requirements. Such a boot loader may use the
1301 following hooks that, if set, are invoked by the kernel at the
1302 appropriate time. The use of these hooks should probably be
1303 considered an absolutely last resort!
1304
1305 IMPORTANT: All the hooks are required to preserve %esp, %ebp, %esi and
1306 %edi across invocation.
1307
1308 realmode_swtch:
1309 A 16-bit real mode far subroutine invoked immediately before
1310 entering protected mode. The default routine disables NMI, so
1311 your routine should probably do so, too.
1312
1313 code32_start:
1314 A 32-bit flat-mode routine *jumped* to immediately after the
1315 transition to protected mode, but before the kernel is
1316 uncompressed. No segments, except CS, are guaranteed to be
1317 set up (current kernels do, but older ones do not); you should
1318 set them up to BOOT_DS (0x18) yourself.
1319
1320 After completing your hook, you should jump to the address
1321 that was in this field before your boot loader overwrote it
1322 (relocated, if appropriate.)
1323
1324
1325 32-bit Boot Protocol
1326 ====================
1327
1328 For machine with some new BIOS other than legacy BIOS, such as EFI,
1329 LinuxBIOS, etc, and kexec, the 16-bit real mode setup code in kernel
1330 based on legacy BIOS can not be used, so a 32-bit boot protocol needs
1331 to be defined.
1332
1333 In 32-bit boot protocol, the first step in loading a Linux kernel
1334 should be to setup the boot parameters (struct boot_params,
1335 traditionally known as "zero page"). The memory for struct boot_params
1336 should be allocated and initialized to all zero. Then the setup header
1337 from offset 0x01f1 of kernel image on should be loaded into struct
1338 boot_params and examined. The end of setup header can be calculated as
1339 follow::
1340
1341 0x0202 + byte value at offset 0x0201
1342
1343 In addition to read/modify/write the setup header of the struct
1344 boot_params as that of 16-bit boot protocol, the boot loader should
1345 also fill the additional fields of the struct boot_params as
1346 described in chapter Documentation/x86/zero-page.rst.
1347
1348 After setting up the struct boot_params, the boot loader can load the
1349 32/64-bit kernel in the same way as that of 16-bit boot protocol.
1350
1351 In 32-bit boot protocol, the kernel is started by jumping to the
1352 32-bit kernel entry point, which is the start address of loaded
1353 32/64-bit kernel.
1354
1355 At entry, the CPU must be in 32-bit protected mode with paging
1356 disabled; a GDT must be loaded with the descriptors for selectors
1357 __BOOT_CS(0x10) and __BOOT_DS(0x18); both descriptors must be 4G flat
1358 segment; __BOOT_CS must have execute/read permission, and __BOOT_DS
1359 must have read/write permission; CS must be __BOOT_CS and DS, ES, SS
1360 must be __BOOT_DS; interrupt must be disabled; %esi must hold the base
1361 address of the struct boot_params; %ebp, %edi and %ebx must be zero.
1362
1363 64-bit Boot Protocol
1364 ====================
1365
1366 For machine with 64bit cpus and 64bit kernel, we could use 64bit bootloader
1367 and we need a 64-bit boot protocol.
1368
1369 In 64-bit boot protocol, the first step in loading a Linux kernel
1370 should be to setup the boot parameters (struct boot_params,
1371 traditionally known as "zero page"). The memory for struct boot_params
1372 could be allocated anywhere (even above 4G) and initialized to all zero.
1373 Then, the setup header at offset 0x01f1 of kernel image on should be
1374 loaded into struct boot_params and examined. The end of setup header
1375 can be calculated as follows::
1376
1377 0x0202 + byte value at offset 0x0201
1378
1379 In addition to read/modify/write the setup header of the struct
1380 boot_params as that of 16-bit boot protocol, the boot loader should
1381 also fill the additional fields of the struct boot_params as described
1382 in chapter Documentation/x86/zero-page.rst.
1383
1384 After setting up the struct boot_params, the boot loader can load
1385 64-bit kernel in the same way as that of 16-bit boot protocol, but
1386 kernel could be loaded above 4G.
1387
1388 In 64-bit boot protocol, the kernel is started by jumping to the
1389 64-bit kernel entry point, which is the start address of loaded
1390 64-bit kernel plus 0x200.
1391
1392 At entry, the CPU must be in 64-bit mode with paging enabled.
1393 The range with setup_header.init_size from start address of loaded
1394 kernel and zero page and command line buffer get ident mapping;
1395 a GDT must be loaded with the descriptors for selectors
1396 __BOOT_CS(0x10) and __BOOT_DS(0x18); both descriptors must be 4G flat
1397 segment; __BOOT_CS must have execute/read permission, and __BOOT_DS
1398 must have read/write permission; CS must be __BOOT_CS and DS, ES, SS
1399 must be __BOOT_DS; interrupt must be disabled; %rsi must hold the base
1400 address of the struct boot_params.
1401
1402 EFI Handover Protocol (deprecated)
1403 ==================================
1404
1405 This protocol allows boot loaders to defer initialisation to the EFI
1406 boot stub. The boot loader is required to load the kernel/initrd(s)
1407 from the boot media and jump to the EFI handover protocol entry point
1408 which is hdr->handover_offset bytes from the beginning of
1409 startup_{32,64}.
1410
1411 The boot loader MUST respect the kernel's PE/COFF metadata when it comes
1412 to section alignment, the memory footprint of the executable image beyond
1413 the size of the file itself, and any other aspect of the PE/COFF header
1414 that may affect correct operation of the image as a PE/COFF binary in the
1415 execution context provided by the EFI firmware.
1416
1417 The function prototype for the handover entry point looks like this::
1418
1419 efi_main(void *handle, efi_system_table_t *table, struct boot_params *bp)
1420
1421 'handle' is the EFI image handle passed to the boot loader by the EFI
1422 firmware, 'table' is the EFI system table - these are the first two
1423 arguments of the "handoff state" as described in section 2.3 of the
1424 UEFI specification. 'bp' is the boot loader-allocated boot params.
1425
1426 The boot loader *must* fill out the following fields in bp::
1427
1428 - hdr.cmd_line_ptr
1429 - hdr.ramdisk_image (if applicable)
1430 - hdr.ramdisk_size (if applicable)
1431
1432 All other fields should be zero.
1433
1434 NOTE: The EFI Handover Protocol is deprecated in favour of the ordinary PE/COFF
1435 entry point, combined with the LINUX_EFI_INITRD_MEDIA_GUID based initrd
1436 loading protocol (refer to [0] for an example of the bootloader side of
1437 this), which removes the need for any knowledge on the part of the EFI
1438 bootloader regarding the internal representation of boot_params or any
1439 requirements/limitations regarding the placement of the command line
1440 and ramdisk in memory, or the placement of the kernel image itself.
1441
1442 [0] https://github.com/u-boot/u-boot/commit/ec80b4735a593961fe701cc3a5d717d4739b0fd0
