1.. SPDX-License-Identifier: GPL-2.0 2 3=========================== 4The Linux/x86 Boot Protocol 5=========================== 6 7On the x86 platform, the Linux kernel uses a rather complicated boot 8convention. This has evolved partially due to historical aspects, as 9well as the desire in the early days to have the kernel itself be a 10bootable image, the complicated PC memory model and due to changed 11expectations in the PC industry caused by the effective demise of 12real-mode DOS as a mainstream operating system. 13 14Currently, the following versions of the Linux/x86 boot protocol exist. 15 16============= ============================================================ 17Old kernels zImage/Image support only. Some very early kernels 18 may not even support a command line. 19 20Protocol 2.00 (Kernel 1.3.73) Added bzImage and initrd support, as 21 well as a formalized way to communicate between the 22 boot loader and the kernel. setup.S made relocatable, 23 although the traditional setup area still assumed 24 writable. 25 26Protocol 2.01 (Kernel 1.3.76) Added a heap overrun warning. 27 28Protocol 2.02 (Kernel 2.4.0-test3-pre3) New command line protocol. 29 Lower the conventional memory ceiling. No overwrite 30 of the traditional setup area, thus making booting 31 safe for systems which use the EBDA from SMM or 32-bit 32 BIOS entry points. zImage deprecated but still 33 supported. 34 35Protocol 2.03 (Kernel 2.4.18-pre1) Explicitly makes the highest possible 36 initrd address available to the bootloader. 37 38Protocol 2.04 (Kernel 2.6.14) Extend the syssize field to four bytes. 39 40Protocol 2.05 (Kernel 2.6.20) Make protected mode kernel relocatable. 41 Introduce relocatable_kernel and kernel_alignment fields. 42 43Protocol 2.06 (Kernel 2.6.22) Added a field that contains the size of 44 the boot command line. 45 46Protocol 2.07 (Kernel 2.6.24) Added paravirtualised boot protocol. 47 Introduced hardware_subarch and hardware_subarch_data 48 and KEEP_SEGMENTS flag in load_flags. 49 50Protocol 2.08 (Kernel 2.6.26) Added crc32 checksum and ELF format 51 payload. Introduced payload_offset and payload_length 52 fields to aid in locating the payload. 53 54Protocol 2.09 (Kernel 2.6.26) Added a field of 64-bit physical 55 pointer to single linked list of struct setup_data. 56 57Protocol 2.10 (Kernel 2.6.31) Added a protocol for relaxed alignment 58 beyond the kernel_alignment added, new init_size and 59 pref_address fields. Added extended boot loader IDs. 60 61Protocol 2.11 (Kernel 3.6) Added a field for offset of EFI handover 62 protocol entry point. 63 64Protocol 2.12 (Kernel 3.8) Added the xloadflags field and extension fields 65 to struct boot_params for loading bzImage and ramdisk 66 above 4G in 64bit. 67 68Protocol 2.13 (Kernel 3.14) Support 32- and 64-bit flags being set in 69 xloadflags to support booting a 64-bit kernel from 32-bit 70 EFI 71 72Protocol 2.14 BURNT BY INCORRECT COMMIT 73 ae7e1238e68f2a472a125673ab506d49158c1889 74 (x86/boot: Add ACPI RSDP address to setup_header) 75 DO NOT USE!!! ASSUME SAME AS 2.13. 76 77Protocol 2.15 (Kernel 5.5) Added the kernel_info and kernel_info.setup_type_max. 78============= ============================================================ 79 80.. note:: 81 The protocol version number should be changed only if the setup header 82 is changed. There is no need to update the version number if boot_params 83 or kernel_info are changed. Additionally, it is recommended to use 84 xloadflags (in this case the protocol version number should not be 85 updated either) or kernel_info to communicate supported Linux kernel 86 features to the boot loader. Due to very limited space available in 87 the original setup header every update to it should be considered 88 with great care. Starting from the protocol 2.15 the primary way to 89 communicate things to the boot loader is the kernel_info. 90 91 92Memory Layout 93============= 94 95The traditional memory map for the kernel loader, used for Image or 96zImage kernels, typically looks like:: 97 98 | | 99 0A0000 +------------------------+ 100 | Reserved for BIOS | Do not use. Reserved for BIOS EBDA. 101 09A000 +------------------------+ 102 | Command line | 103 | Stack/heap | For use by the kernel real-mode code. 104 098000 +------------------------+ 105 | Kernel setup | The kernel real-mode code. 106 090200 +------------------------+ 107 | Kernel boot sector | The kernel legacy boot sector. 108 090000 +------------------------+ 109 | Protected-mode kernel | The bulk of the kernel image. 110 010000 +------------------------+ 111 | Boot loader | <- Boot sector entry point 0000:7C00 112 001000 +------------------------+ 113 | Reserved for MBR/BIOS | 114 000800 +------------------------+ 115 | Typically used by MBR | 116 000600 +------------------------+ 117 | BIOS use only | 118 000000 +------------------------+ 119 120When using bzImage, the protected-mode kernel was relocated to 1210x100000 ("high memory"), and the kernel real-mode block (boot sector, 122setup, and stack/heap) was made relocatable to any address between 1230x10000 and end of low memory. Unfortunately, in protocols 2.00 and 1242.01 the 0x90000+ memory range is still used internally by the kernel; 125the 2.02 protocol resolves that problem. 126 127It is desirable to keep the "memory ceiling" -- the highest point in 128low memory touched by the boot loader -- as low as possible, since 129some newer BIOSes have begun to allocate some rather large amounts of 130memory, called the Extended BIOS Data Area, near the top of low 131memory. The boot loader should use the "INT 12h" BIOS call to verify 132how much low memory is available. 133 134Unfortunately, if INT 12h reports that the amount of memory is too 135low, there is usually nothing the boot loader can do but to report an 136error to the user. The boot loader should therefore be designed to 137take up as little space in low memory as it reasonably can. For 138zImage or old bzImage kernels, which need data written into the 1390x90000 segment, the boot loader should make sure not to use memory 140above the 0x9A000 point; too many BIOSes will break above that point. 141 142For a modern bzImage kernel with boot protocol version >= 2.02, a 143memory layout like the following is suggested:: 144 145 ~ ~ 146 | Protected-mode kernel | 147 100000 +------------------------+ 148 | I/O memory hole | 149 0A0000 +------------------------+ 150 | Reserved for BIOS | Leave as much as possible unused 151 ~ ~ 152 | Command line | (Can also be below the X+10000 mark) 153 X+10000 +------------------------+ 154 | Stack/heap | For use by the kernel real-mode code. 155 X+08000 +------------------------+ 156 | Kernel setup | The kernel real-mode code. 157 | Kernel boot sector | The kernel legacy boot sector. 158 X +------------------------+ 159 | Boot loader | <- Boot sector entry point 0000:7C00 160 001000 +------------------------+ 161 | Reserved for MBR/BIOS | 162 000800 +------------------------+ 163 | Typically used by MBR | 164 000600 +------------------------+ 165 | BIOS use only | 166 000000 +------------------------+ 167 168 ... where the address X is as low as the design of the boot loader permits. 169 170 171The Real-Mode Kernel Header 172=========================== 173 174In the following text, and anywhere in the kernel boot sequence, "a 175sector" refers to 512 bytes. It is independent of the actual sector 176size of the underlying medium. 177 178The first step in loading a Linux kernel should be to load the 179real-mode code (boot sector and setup code) and then examine the 180following header at offset 0x01f1. The real-mode code can total up to 18132K, although the boot loader may choose to load only the first two 182sectors (1K) and then examine the bootup sector size. 183 184The header looks like: 185 186=========== ======== ===================== ============================================ 187Offset/Size Proto Name Meaning 188=========== ======== ===================== ============================================ 18901F1/1 ALL(1) setup_sects The size of the setup in sectors 19001F2/2 ALL root_flags If set, the root is mounted readonly 19101F4/4 2.04+(2) syssize The size of the 32-bit code in 16-byte paras 19201F8/2 ALL ram_size DO NOT USE - for bootsect.S use only 19301FA/2 ALL vid_mode Video mode control 19401FC/2 ALL root_dev Default root device number 19501FE/2 ALL boot_flag 0xAA55 magic number 1960200/2 2.00+ jump Jump instruction 1970202/4 2.00+ header Magic signature "HdrS" 1980206/2 2.00+ version Boot protocol version supported 1990208/4 2.00+ realmode_swtch Boot loader hook (see below) 200020C/2 2.00+ start_sys_seg The load-low segment (0x1000) (obsolete) 201020E/2 2.00+ kernel_version Pointer to kernel version string 2020210/1 2.00+ type_of_loader Boot loader identifier 2030211/1 2.00+ loadflags Boot protocol option flags 2040212/2 2.00+ setup_move_size Move to high memory size (used with hooks) 2050214/4 2.00+ code32_start Boot loader hook (see below) 2060218/4 2.00+ ramdisk_image initrd load address (set by boot loader) 207021C/4 2.00+ ramdisk_size initrd size (set by boot loader) 2080220/4 2.00+ bootsect_kludge DO NOT USE - for bootsect.S use only 2090224/2 2.01+ heap_end_ptr Free memory after setup end 2100226/1 2.02+(3) ext_loader_ver Extended boot loader version 2110227/1 2.02+(3) ext_loader_type Extended boot loader ID 2120228/4 2.02+ cmd_line_ptr 32-bit pointer to the kernel command line 213022C/4 2.03+ initrd_addr_max Highest legal initrd address 2140230/4 2.05+ kernel_alignment Physical addr alignment required for kernel 2150234/1 2.05+ relocatable_kernel Whether kernel is relocatable or not 2160235/1 2.10+ min_alignment Minimum alignment, as a power of two 2170236/2 2.12+ xloadflags Boot protocol option flags 2180238/4 2.06+ cmdline_size Maximum size of the kernel command line 219023C/4 2.07+ hardware_subarch Hardware subarchitecture 2200240/8 2.07+ hardware_subarch_data Subarchitecture-specific data 2210248/4 2.08+ payload_offset Offset of kernel payload 222024C/4 2.08+ payload_length Length of kernel payload 2230250/8 2.09+ setup_data 64-bit physical pointer to linked list 224 of struct setup_data 2250258/8 2.10+ pref_address Preferred loading address 2260260/4 2.10+ init_size Linear memory required during initialization 2270264/4 2.11+ handover_offset Offset of handover entry point 2280268/4 2.15+ kernel_info_offset Offset of the kernel_info 229=========== ======== ===================== ============================================ 230 231.. note:: 232 (1) For backwards compatibility, if the setup_sects field contains 0, the 233 real value is 4. 234 235 (2) For boot protocol prior to 2.04, the upper two bytes of the syssize 236 field are unusable, which means the size of a bzImage kernel 237 cannot be determined. 238 239 (3) Ignored, but safe to set, for boot protocols 2.02-2.09. 240 241If the "HdrS" (0x53726448) magic number is not found at offset 0x202, 242the boot protocol version is "old". Loading an old kernel, the 243following parameters should be assumed:: 244 245 Image type = zImage 246 initrd not supported 247 Real-mode kernel must be located at 0x90000. 248 249Otherwise, the "version" field contains the protocol version, 250e.g. protocol version 2.01 will contain 0x0201 in this field. When 251setting fields in the header, you must make sure only to set fields 252supported by the protocol version in use. 253 254 255Details of Header Fields 256======================== 257 258For each field, some are information from the kernel to the bootloader 259("read"), some are expected to be filled out by the bootloader 260("write"), and some are expected to be read and modified by the 261bootloader ("modify"). 262 263All general purpose boot loaders should write the fields marked 264(obligatory). Boot loaders who want to load the kernel at a 265nonstandard address should fill in the fields marked (reloc); other 266boot loaders can ignore those fields. 267 268The byte order of all fields is littleendian (this is x86, after all.) 269 270============ =========== 271Field name: setup_sects 272Type: read 273Offset/size: 0x1f1/1 274Protocol: ALL 275============ =========== 276 277 The size of the setup code in 512-byte sectors. If this field is 278 0, the real value is 4. The real-mode code consists of the boot 279 sector (always one 512-byte sector) plus the setup code. 280 281============ ================= 282Field name: root_flags 283Type: modify (optional) 284Offset/size: 0x1f2/2 285Protocol: ALL 286============ ================= 287 288 If this field is nonzero, the root defaults to readonly. The use of 289 this field is deprecated; use the "ro" or "rw" options on the 290 command line instead. 291 292============ =============================================== 293Field name: syssize 294Type: read 295Offset/size: 0x1f4/4 (protocol 2.04+) 0x1f4/2 (protocol ALL) 296Protocol: 2.04+ 297============ =============================================== 298 299 The size of the protected-mode code in units of 16-byte paragraphs. 300 For protocol versions older than 2.04 this field is only two bytes 301 wide, and therefore cannot be trusted for the size of a kernel if 302 the LOAD_HIGH flag is set. 303 304============ =============== 305Field name: ram_size 306Type: kernel internal 307Offset/size: 0x1f8/2 308Protocol: ALL 309============ =============== 310 311 This field is obsolete. 312 313============ =================== 314Field name: vid_mode 315Type: modify (obligatory) 316Offset/size: 0x1fa/2 317============ =================== 318 319 Please see the section on SPECIAL COMMAND LINE OPTIONS. 320 321============ ================= 322Field name: root_dev 323Type: modify (optional) 324Offset/size: 0x1fc/2 325Protocol: ALL 326============ ================= 327 328 The default root device device number. The use of this field is 329 deprecated, use the "root=" option on the command line instead. 330 331============ ========= 332Field name: boot_flag 333Type: read 334Offset/size: 0x1fe/2 335Protocol: ALL 336============ ========= 337 338 Contains 0xAA55. This is the closest thing old Linux kernels have 339 to a magic number. 340 341============ ======= 342Field name: jump 343Type: read 344Offset/size: 0x200/2 345Protocol: 2.00+ 346============ ======= 347 348 Contains an x86 jump instruction, 0xEB followed by a signed offset 349 relative to byte 0x202. This can be used to determine the size of 350 the header. 351 352============ ======= 353Field name: header 354Type: read 355Offset/size: 0x202/4 356Protocol: 2.00+ 357============ ======= 358 359 Contains the magic number "HdrS" (0x53726448). 360 361============ ======= 362Field name: version 363Type: read 364Offset/size: 0x206/2 365Protocol: 2.00+ 366============ ======= 367 368 Contains the boot protocol version, in (major << 8)+minor format, 369 e.g. 0x0204 for version 2.04, and 0x0a11 for a hypothetical version 370 10.17. 371 372============ ================= 373Field name: realmode_swtch 374Type: modify (optional) 375Offset/size: 0x208/4 376Protocol: 2.00+ 377============ ================= 378 379 Boot loader hook (see ADVANCED BOOT LOADER HOOKS below.) 380 381============ ============= 382Field name: start_sys_seg 383Type: read 384Offset/size: 0x20c/2 385Protocol: 2.00+ 386============ ============= 387 388 The load low segment (0x1000). Obsolete. 389 390============ ============== 391Field name: kernel_version 392Type: read 393Offset/size: 0x20e/2 394Protocol: 2.00+ 395============ ============== 396 397 If set to a nonzero value, contains a pointer to a NUL-terminated 398 human-readable kernel version number string, less 0x200. This can 399 be used to display the kernel version to the user. This value 400 should be less than (0x200*setup_sects). 401 402 For example, if this value is set to 0x1c00, the kernel version 403 number string can be found at offset 0x1e00 in the kernel file. 404 This is a valid value if and only if the "setup_sects" field 405 contains the value 15 or higher, as:: 406 407 0x1c00 < 15*0x200 (= 0x1e00) but 408 0x1c00 >= 14*0x200 (= 0x1c00) 409 410 0x1c00 >> 9 = 14, So the minimum value for setup_secs is 15. 411 412============ ================== 413Field name: type_of_loader 414Type: write (obligatory) 415Offset/size: 0x210/1 416Protocol: 2.00+ 417============ ================== 418 419 If your boot loader has an assigned id (see table below), enter 420 0xTV here, where T is an identifier for the boot loader and V is 421 a version number. Otherwise, enter 0xFF here. 422 423 For boot loader IDs above T = 0xD, write T = 0xE to this field and 424 write the extended ID minus 0x10 to the ext_loader_type field. 425 Similarly, the ext_loader_ver field can be used to provide more than 426 four bits for the bootloader version. 427 428 For example, for T = 0x15, V = 0x234, write:: 429 430 type_of_loader <- 0xE4 431 ext_loader_type <- 0x05 432 ext_loader_ver <- 0x23 433 434 Assigned boot loader ids (hexadecimal): 435 436 == ======================================= 437 0 LILO 438 (0x00 reserved for pre-2.00 bootloader) 439 1 Loadlin 440 2 bootsect-loader 441 (0x20, all other values reserved) 442 3 Syslinux 443 4 Etherboot/gPXE/iPXE 444 5 ELILO 445 7 GRUB 446 8 U-Boot 447 9 Xen 448 A Gujin 449 B Qemu 450 C Arcturus Networks uCbootloader 451 D kexec-tools 452 E Extended (see ext_loader_type) 453 F Special (0xFF = undefined) 454 10 Reserved 455 11 Minimal Linux Bootloader 456 <http://sebastian-plotz.blogspot.de> 457 12 OVMF UEFI virtualization stack 458 == ======================================= 459 460 Please contact <hpa@zytor.com> if you need a bootloader ID value assigned. 461 462============ =================== 463Field name: loadflags 464Type: modify (obligatory) 465Offset/size: 0x211/1 466Protocol: 2.00+ 467============ =================== 468 469 This field is a bitmask. 470 471 Bit 0 (read): LOADED_HIGH 472 473 - If 0, the protected-mode code is loaded at 0x10000. 474 - If 1, the protected-mode code is loaded at 0x100000. 475 476 Bit 1 (kernel internal): KASLR_FLAG 477 478 - Used internally by the compressed kernel to communicate 479 KASLR status to kernel proper. 480 481 - If 1, KASLR enabled. 482 - If 0, KASLR disabled. 483 484 Bit 5 (write): QUIET_FLAG 485 486 - If 0, print early messages. 487 - If 1, suppress early messages. 488 489 This requests to the kernel (decompressor and early 490 kernel) to not write early messages that require 491 accessing the display hardware directly. 492 493 Bit 6 (obsolete): KEEP_SEGMENTS 494 495 Protocol: 2.07+ 496 497 - This flag is obsolete. 498 499 Bit 7 (write): CAN_USE_HEAP 500 501 Set this bit to 1 to indicate that the value entered in the 502 heap_end_ptr is valid. If this field is clear, some setup code 503 functionality will be disabled. 504 505 506============ =================== 507Field name: setup_move_size 508Type: modify (obligatory) 509Offset/size: 0x212/2 510Protocol: 2.00-2.01 511============ =================== 512 513 When using protocol 2.00 or 2.01, if the real mode kernel is not 514 loaded at 0x90000, it gets moved there later in the loading 515 sequence. Fill in this field if you want additional data (such as 516 the kernel command line) moved in addition to the real-mode kernel 517 itself. 518 519 The unit is bytes starting with the beginning of the boot sector. 520 521 This field is can be ignored when the protocol is 2.02 or higher, or 522 if the real-mode code is loaded at 0x90000. 523 524============ ======================== 525Field name: code32_start 526Type: modify (optional, reloc) 527Offset/size: 0x214/4 528Protocol: 2.00+ 529============ ======================== 530 531 The address to jump to in protected mode. This defaults to the load 532 address of the kernel, and can be used by the boot loader to 533 determine the proper load address. 534 535 This field can be modified for two purposes: 536 537 1. as a boot loader hook (see Advanced Boot Loader Hooks below.) 538 539 2. if a bootloader which does not install a hook loads a 540 relocatable kernel at a nonstandard address it will have to modify 541 this field to point to the load address. 542 543============ ================== 544Field name: ramdisk_image 545Type: write (obligatory) 546Offset/size: 0x218/4 547Protocol: 2.00+ 548============ ================== 549 550 The 32-bit linear address of the initial ramdisk or ramfs. Leave at 551 zero if there is no initial ramdisk/ramfs. 552 553============ ================== 554Field name: ramdisk_size 555Type: write (obligatory) 556Offset/size: 0x21c/4 557Protocol: 2.00+ 558============ ================== 559 560 Size of the initial ramdisk or ramfs. Leave at zero if there is no 561 initial ramdisk/ramfs. 562 563============ =============== 564Field name: bootsect_kludge 565Type: kernel internal 566Offset/size: 0x220/4 567Protocol: 2.00+ 568============ =============== 569 570 This field is obsolete. 571 572============ ================== 573Field name: heap_end_ptr 574Type: write (obligatory) 575Offset/size: 0x224/2 576Protocol: 2.01+ 577============ ================== 578 579 Set this field to the offset (from the beginning of the real-mode 580 code) of the end of the setup stack/heap, minus 0x0200. 581 582============ ================ 583Field name: ext_loader_ver 584Type: write (optional) 585Offset/size: 0x226/1 586Protocol: 2.02+ 587============ ================ 588 589 This field is used as an extension of the version number in the 590 type_of_loader field. The total version number is considered to be 591 (type_of_loader & 0x0f) + (ext_loader_ver << 4). 592 593 The use of this field is boot loader specific. If not written, it 594 is zero. 595 596 Kernels prior to 2.6.31 did not recognize this field, but it is safe 597 to write for protocol version 2.02 or higher. 598 599============ ===================================================== 600Field name: ext_loader_type 601Type: write (obligatory if (type_of_loader & 0xf0) == 0xe0) 602Offset/size: 0x227/1 603Protocol: 2.02+ 604============ ===================================================== 605 606 This field is used as an extension of the type number in 607 type_of_loader field. If the type in type_of_loader is 0xE, then 608 the actual type is (ext_loader_type + 0x10). 609 610 This field is ignored if the type in type_of_loader is not 0xE. 611 612 Kernels prior to 2.6.31 did not recognize this field, but it is safe 613 to write for protocol version 2.02 or higher. 614 615============ ================== 616Field name: cmd_line_ptr 617Type: write (obligatory) 618Offset/size: 0x228/4 619Protocol: 2.02+ 620============ ================== 621 622 Set this field to the linear address of the kernel command line. 623 The kernel command line can be located anywhere between the end of 624 the setup heap and 0xA0000; it does not have to be located in the 625 same 64K segment as the real-mode code itself. 626 627 Fill in this field even if your boot loader does not support a 628 command line, in which case you can point this to an empty string 629 (or better yet, to the string "auto".) If this field is left at 630 zero, the kernel will assume that your boot loader does not support 631 the 2.02+ protocol. 632 633============ =============== 634Field name: initrd_addr_max 635Type: read 636Offset/size: 0x22c/4 637Protocol: 2.03+ 638============ =============== 639 640 The maximum address that may be occupied by the initial 641 ramdisk/ramfs contents. For boot protocols 2.02 or earlier, this 642 field is not present, and the maximum address is 0x37FFFFFF. (This 643 address is defined as the address of the highest safe byte, so if 644 your ramdisk is exactly 131072 bytes long and this field is 645 0x37FFFFFF, you can start your ramdisk at 0x37FE0000.) 646 647============ ============================ 648Field name: kernel_alignment 649Type: read/modify (reloc) 650Offset/size: 0x230/4 651Protocol: 2.05+ (read), 2.10+ (modify) 652============ ============================ 653 654 Alignment unit required by the kernel (if relocatable_kernel is 655 true.) A relocatable kernel that is loaded at an alignment 656 incompatible with the value in this field will be realigned during 657 kernel initialization. 658 659 Starting with protocol version 2.10, this reflects the kernel 660 alignment preferred for optimal performance; it is possible for the 661 loader to modify this field to permit a lesser alignment. See the 662 min_alignment and pref_address field below. 663 664============ ================== 665Field name: relocatable_kernel 666Type: read (reloc) 667Offset/size: 0x234/1 668Protocol: 2.05+ 669============ ================== 670 671 If this field is nonzero, the protected-mode part of the kernel can 672 be loaded at any address that satisfies the kernel_alignment field. 673 After loading, the boot loader must set the code32_start field to 674 point to the loaded code, or to a boot loader hook. 675 676============ ============= 677Field name: min_alignment 678Type: read (reloc) 679Offset/size: 0x235/1 680Protocol: 2.10+ 681============ ============= 682 683 This field, if nonzero, indicates as a power of two the minimum 684 alignment required, as opposed to preferred, by the kernel to boot. 685 If a boot loader makes use of this field, it should update the 686 kernel_alignment field with the alignment unit desired; typically:: 687 688 kernel_alignment = 1 << min_alignment 689 690 There may be a considerable performance cost with an excessively 691 misaligned kernel. Therefore, a loader should typically try each 692 power-of-two alignment from kernel_alignment down to this alignment. 693 694============ ========== 695Field name: xloadflags 696Type: read 697Offset/size: 0x236/2 698Protocol: 2.12+ 699============ ========== 700 701 This field is a bitmask. 702 703 Bit 0 (read): XLF_KERNEL_64 704 705 - If 1, this kernel has the legacy 64-bit entry point at 0x200. 706 707 Bit 1 (read): XLF_CAN_BE_LOADED_ABOVE_4G 708 709 - If 1, kernel/boot_params/cmdline/ramdisk can be above 4G. 710 711 Bit 2 (read): XLF_EFI_HANDOVER_32 712 713 - If 1, the kernel supports the 32-bit EFI handoff entry point 714 given at handover_offset. 715 716 Bit 3 (read): XLF_EFI_HANDOVER_64 717 718 - If 1, the kernel supports the 64-bit EFI handoff entry point 719 given at handover_offset + 0x200. 720 721 Bit 4 (read): XLF_EFI_KEXEC 722 723 - If 1, the kernel supports kexec EFI boot with EFI runtime support. 724 725 726============ ============ 727Field name: cmdline_size 728Type: read 729Offset/size: 0x238/4 730Protocol: 2.06+ 731============ ============ 732 733 The maximum size of the command line without the terminating 734 zero. This means that the command line can contain at most 735 cmdline_size characters. With protocol version 2.05 and earlier, the 736 maximum size was 255. 737 738============ ==================================== 739Field name: hardware_subarch 740Type: write (optional, defaults to x86/PC) 741Offset/size: 0x23c/4 742Protocol: 2.07+ 743============ ==================================== 744 745 In a paravirtualized environment the hardware low level architectural 746 pieces such as interrupt handling, page table handling, and 747 accessing process control registers needs to be done differently. 748 749 This field allows the bootloader to inform the kernel we are in one 750 one of those environments. 751 752 ========== ============================== 753 0x00000000 The default x86/PC environment 754 0x00000001 lguest 755 0x00000002 Xen 756 0x00000003 Moorestown MID 757 0x00000004 CE4100 TV Platform 758 ========== ============================== 759 760============ ========================= 761Field name: hardware_subarch_data 762Type: write (subarch-dependent) 763Offset/size: 0x240/8 764Protocol: 2.07+ 765============ ========================= 766 767 A pointer to data that is specific to hardware subarch 768 This field is currently unused for the default x86/PC environment, 769 do not modify. 770 771============ ============== 772Field name: payload_offset 773Type: read 774Offset/size: 0x248/4 775Protocol: 2.08+ 776============ ============== 777 778 If non-zero then this field contains the offset from the beginning 779 of the protected-mode code to the payload. 780 781 The payload may be compressed. The format of both the compressed and 782 uncompressed data should be determined using the standard magic 783 numbers. The currently supported compression formats are gzip 784 (magic numbers 1F 8B or 1F 9E), bzip2 (magic number 42 5A), LZMA 785 (magic number 5D 00), XZ (magic number FD 37), LZ4 (magic number 786 02 21) and ZSTD (magic number 28 B5). The uncompressed payload is 787 currently always ELF (magic number 7F 45 4C 46). 788 789============ ============== 790Field name: payload_length 791Type: read 792Offset/size: 0x24c/4 793Protocol: 2.08+ 794============ ============== 795 796 The length of the payload. 797 798============ =============== 799Field name: setup_data 800Type: write (special) 801Offset/size: 0x250/8 802Protocol: 2.09+ 803============ =============== 804 805 The 64-bit physical pointer to NULL terminated single linked list of 806 struct setup_data. This is used to define a more extensible boot 807 parameters passing mechanism. The definition of struct setup_data is 808 as follow:: 809 810 struct setup_data { 811 u64 next; 812 u32 type; 813 u32 len; 814 u8 data[0]; 815 }; 816 817 Where, the next is a 64-bit physical pointer to the next node of 818 linked list, the next field of the last node is 0; the type is used 819 to identify the contents of data; the len is the length of data 820 field; the data holds the real payload. 821 822 This list may be modified at a number of points during the bootup 823 process. Therefore, when modifying this list one should always make 824 sure to consider the case where the linked list already contains 825 entries. 826 827 The setup_data is a bit awkward to use for extremely large data objects, 828 both because the setup_data header has to be adjacent to the data object 829 and because it has a 32-bit length field. However, it is important that 830 intermediate stages of the boot process have a way to identify which 831 chunks of memory are occupied by kernel data. 832 833 Thus setup_indirect struct and SETUP_INDIRECT type were introduced in 834 protocol 2.15:: 835 836 struct setup_indirect { 837 __u32 type; 838 __u32 reserved; /* Reserved, must be set to zero. */ 839 __u64 len; 840 __u64 addr; 841 }; 842 843 The type member is a SETUP_INDIRECT | SETUP_* type. However, it cannot be 844 SETUP_INDIRECT itself since making the setup_indirect a tree structure 845 could require a lot of stack space in something that needs to parse it 846 and stack space can be limited in boot contexts. 847 848 Let's give an example how to point to SETUP_E820_EXT data using setup_indirect. 849 In this case setup_data and setup_indirect will look like this:: 850 851 struct setup_data { 852 __u64 next = 0 or <addr_of_next_setup_data_struct>; 853 __u32 type = SETUP_INDIRECT; 854 __u32 len = sizeof(setup_indirect); 855 __u8 data[sizeof(setup_indirect)] = struct setup_indirect { 856 __u32 type = SETUP_INDIRECT | SETUP_E820_EXT; 857 __u32 reserved = 0; 858 __u64 len = <len_of_SETUP_E820_EXT_data>; 859 __u64 addr = <addr_of_SETUP_E820_EXT_data>; 860 } 861 } 862 863.. note:: 864 SETUP_INDIRECT | SETUP_NONE objects cannot be properly distinguished 865 from SETUP_INDIRECT itself. So, this kind of objects cannot be provided 866 by the bootloaders. 867 868============ ============ 869Field name: pref_address 870Type: read (reloc) 871Offset/size: 0x258/8 872Protocol: 2.10+ 873============ ============ 874 875 This field, if nonzero, represents a preferred load address for the 876 kernel. A relocating bootloader should attempt to load at this 877 address if possible. 878 879 A non-relocatable kernel will unconditionally move itself and to run 880 at this address. 881 882============ ======= 883Field name: init_size 884Type: read 885Offset/size: 0x260/4 886============ ======= 887 888 This field indicates the amount of linear contiguous memory starting 889 at the kernel runtime start address that the kernel needs before it 890 is capable of examining its memory map. This is not the same thing 891 as the total amount of memory the kernel needs to boot, but it can 892 be used by a relocating boot loader to help select a safe load 893 address for the kernel. 894 895 The kernel runtime start address is determined by the following algorithm:: 896 897 if (relocatable_kernel) 898 runtime_start = align_up(load_address, kernel_alignment) 899 else 900 runtime_start = pref_address 901 902============ =============== 903Field name: handover_offset 904Type: read 905Offset/size: 0x264/4 906============ =============== 907 908 This field is the offset from the beginning of the kernel image to 909 the EFI handover protocol entry point. Boot loaders using the EFI 910 handover protocol to boot the kernel should jump to this offset. 911 912 See EFI HANDOVER PROTOCOL below for more details. 913 914============ ================== 915Field name: kernel_info_offset 916Type: read 917Offset/size: 0x268/4 918Protocol: 2.15+ 919============ ================== 920 921 This field is the offset from the beginning of the kernel image to the 922 kernel_info. The kernel_info structure is embedded in the Linux image 923 in the uncompressed protected mode region. 924 925 926The kernel_info 927=============== 928 929The relationships between the headers are analogous to the various data 930sections: 931 932 setup_header = .data 933 boot_params/setup_data = .bss 934 935What is missing from the above list? That's right: 936 937 kernel_info = .rodata 938 939We have been (ab)using .data for things that could go into .rodata or .bss for 940a long time, for lack of alternatives and -- especially early on -- inertia. 941Also, the BIOS stub is responsible for creating boot_params, so it isn't 942available to a BIOS-based loader (setup_data is, though). 943 944setup_header is permanently limited to 144 bytes due to the reach of the 9452-byte jump field, which doubles as a length field for the structure, combined 946with the size of the "hole" in struct boot_params that a protected-mode loader 947or the BIOS stub has to copy it into. It is currently 119 bytes long, which 948leaves us with 25 very precious bytes. This isn't something that can be fixed 949without revising the boot protocol entirely, breaking backwards compatibility. 950 951boot_params proper is limited to 4096 bytes, but can be arbitrarily extended 952by adding setup_data entries. It cannot be used to communicate properties of 953the kernel image, because it is .bss and has no image-provided content. 954 955kernel_info solves this by providing an extensible place for information about 956the kernel image. It is readonly, because the kernel cannot rely on a 957bootloader copying its contents anywhere, but that is OK; if it becomes 958necessary it can still contain data items that an enabled bootloader would be 959expected to copy into a setup_data chunk. 960 961All kernel_info data should be part of this structure. Fixed size data have to 962be put before kernel_info_var_len_data label. Variable size data have to be put 963after kernel_info_var_len_data label. Each chunk of variable size data has to 964be prefixed with header/magic and its size, e.g.:: 965 966 kernel_info: 967 .ascii "LToP" /* Header, Linux top (structure). */ 968 .long kernel_info_var_len_data - kernel_info 969 .long kernel_info_end - kernel_info 970 .long 0x01234567 /* Some fixed size data for the bootloaders. */ 971 kernel_info_var_len_data: 972 example_struct: /* Some variable size data for the bootloaders. */ 973 .ascii "0123" /* Header/Magic. */ 974 .long example_struct_end - example_struct 975 .ascii "Struct" 976 .long 0x89012345 977 example_struct_end: 978 example_strings: /* Some variable size data for the bootloaders. */ 979 .ascii "ABCD" /* Header/Magic. */ 980 .long example_strings_end - example_strings 981 .asciz "String_0" 982 .asciz "String_1" 983 example_strings_end: 984 kernel_info_end: 985 986This way the kernel_info is self-contained blob. 987 988.. note:: 989 Each variable size data header/magic can be any 4-character string, 990 without \0 at the end of the string, which does not collide with 991 existing variable length data headers/magics. 992 993 994Details of the kernel_info Fields 995================================= 996 997============ ======== 998Field name: header 999Offset/size: 0x0000/4 1000============ ========
1001 1002 Contains the magic number "LToP" (0x506f544c). 1003 1004============ ======== 1005Field name: size 1006Offset/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============ ======== 1015Field name: size_total 1016Offset/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============ ============== 1023Field name: setup_type_max 1024Offset/size: 0x000c/4 1025============ ============== 1026 1027 This field contains maximal allowed type for setup_data and setup_indirect structs. 1028 1029 1030The Image Checksum 1031================== 1032 1033From boot protocol version 2.08 onwards the CRC-32 is calculated over 1034the entire file using the characteristic polynomial 0x04C11DB7 and an 1035initial remainder of 0xffffffff. The checksum is appended to the 1036file; therefore the CRC of the file up to the limit specified in the 1037syssize field of the header is always 0. 1038 1039 1040The Kernel Command Line 1041======================= 1042 1043The kernel command line has become an important way for the boot 1044loader to communicate with the kernel. Some of its options are also 1045relevant to the boot loader itself, see "special command line options" 1046below. 1047 1048The kernel command line is a null-terminated string. The maximum 1049length can be retrieved from the field cmdline_size. Before protocol 1050version 2.06, the maximum was 255 characters. A string that is too 1051long will be automatically truncated by the kernel. 1052 1053If the boot protocol version is 2.02 or later, the address of the 1054kernel command line is given by the header field cmd_line_ptr (see 1055above.) This address can be anywhere between the end of the setup 1056heap and 0xA0000. 1057 1058If the protocol version is *not* 2.02 or higher, the kernel 1059command 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 1073Memory Layout of The Real-Mode Code 1074=================================== 1075 1076The real-mode code requires a stack/heap to be set up, as well as 1077memory allocated for the kernel command line. This needs to be done 1078in the real-mode accessible memory in bottom megabyte. 1079 1080It should be noted that modern machines often have a sizable Extended 1081BIOS Data Area (EBDA). As a result, it is advisable to use as little 1082of the low megabyte as possible. 1083 1084Unfortunately, under the following circumstances the 0x90000 memory 1085segment 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 1096When loading at 0x90000, avoid using memory above 0x9a000. 1097 1098For boot protocol 2.02 or higher, the command line does not have to be 1099located in the same 64K segment as the real-mode setup code; it is 1100thus permitted to give the stack/heap the full 64K segment and locate 1101the command line above it. 1102 1103The kernel command line should not be located below the real-mode 1104code, nor should it be located in high memory. 1105 1106 1107Sample Boot Configuartion 1108========================= 1109 1110As a sample configuration, assume the following layout of the real 1111mode 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 1129Such 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 1188Loading The Rest of The Kernel 1189============================== 1190 1191The 32-bit (non-real-mode) kernel starts at offset (setup_sects+1)*512 1192in the kernel file (again, if setup_sects == 0 the real value is 4.) 1193It should be loaded at address 0x10000 for Image/zImage kernels and 11940x100000 for bzImage kernels. 1195 1196The kernel is a bzImage kernel if the protocol >= 2.00 and the 0x01 1197bit (LOAD_HIGH) in the loadflags field is set:: 1198 1199 is_bzImage = (protocol >= 0x0200) && (loadflags & 0x01); 1200 load_address = is_bzImage ? 0x100000 : 0x10000; 1201 1202Note that Image/zImage kernels can be up to 512K in size, and thus use 1203the entire 0x10000-0x90000 range of memory. This means it is pretty 1204much a requirement for these kernels to load the real-mode part at 12050x90000. bzImage kernels allow much more flexibility. 1206 1207Special Command Line Options 1208============================ 1209 1210If the command line provided by the boot loader is entered by the 1211user, the user may expect the following command line options to work. 1212They should normally not be deleted from the kernel command line even 1213though not all of them are actually meaningful to the kernel. Boot 1214loader authors who need additional command line options for the boot 1215loader itself should get them registered in 1216Documentation/admin-guide/kernel-parameters.rst to make sure they will not 1217conflict 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 1241In addition, some boot loaders add the following options to the 1242user-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 1251If these options are added by the boot loader, it is highly 1252recommended that they are located *first*, before the user-specified 1253or configuration-specified command line. Otherwise, "init=/bin/sh" 1254gets confused by the "auto" option. 1255 1256 1257Running the Kernel 1258================== 1259 1260The kernel is started by jumping to the kernel entry point, which is 1261located at *segment* offset 0x20 from the start of the real mode 1262kernel. This means that if you loaded your real-mode kernel code at 12630x90000, the kernel entry point is 9020:0000. 1264 1265At entry, ds = es = ss should point to the start of the real-mode 1266kernel code (0x9000 if the code is loaded at 0x90000), sp should be 1267set up properly, normally pointing to the top of the heap, and 1268interrupts should be disabled. Furthermore, to guard against bugs in 1269the kernel, it is recommended that the boot loader sets fs = gs = ds = 1270es = ss. 1271 1272In 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 1288If your boot sector accesses a floppy drive, it is recommended to 1289switch off the floppy motor before running the kernel, since the 1290kernel boot leaves interrupts off and thus the motor will not be 1291switched off, especially if the loaded kernel has the floppy driver as 1292a demand-loaded module! 1293 1294 1295Advanced Boot Loader Hooks 1296========================== 1297 1298If the boot loader runs in a particularly hostile environment (such as 1299LOADLIN, which runs under DOS) it may be impossible to follow the 1300standard memory location requirements. Such a boot loader may use the 1301following hooks that, if set, are invoked by the kernel at the 1302appropriate time. The use of these hooks should probably be 1303considered an absolutely last resort! 1304 1305IMPORTANT: 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 132532-bit Boot Protocol 1326==================== 1327 1328For machine with some new BIOS other than legacy BIOS, such as EFI, 1329LinuxBIOS, etc, and kexec, the 16-bit real mode setup code in kernel 1330based on legacy BIOS can not be used, so a 32-bit boot protocol needs 1331to be defined. 1332 1333In 32-bit boot protocol, the first step in loading a Linux kernel 1334should be to setup the boot parameters (struct boot_params, 1335traditionally known as "zero page"). The memory for struct boot_params 1336should be allocated and initialized to all zero. Then the setup header 1337from offset 0x01f1 of kernel image on should be loaded into struct 1338boot_params and examined. The end of setup header can be calculated as 1339follow:: 1340 1341 0x0202 + byte value at offset 0x0201 1342 1343In addition to read/modify/write the setup header of the struct 1344boot_params as that of 16-bit boot protocol, the boot loader should 1345also fill the additional fields of the struct boot_params as 1346described in chapter Documentation/x86/zero-page.rst. 1347 1348After setting up the struct boot_params, the boot loader can load the 134932/64-bit kernel in the same way as that of 16-bit boot protocol. 1350 1351In 32-bit boot protocol, the kernel is started by jumping to the 135232-bit kernel entry point, which is the start address of loaded 135332/64-bit kernel. 1354 1355At entry, the CPU must be in 32-bit protected mode with paging 1356disabled; a GDT must be loaded with the descriptors for selectors 1357__BOOT_CS(0x10) and __BOOT_DS(0x18); both descriptors must be 4G flat 1358segment; __BOOT_CS must have execute/read permission, and __BOOT_DS 1359must have read/write permission; CS must be __BOOT_CS and DS, ES, SS 1360must be __BOOT_DS; interrupt must be disabled; %esi must hold the base 1361address of the struct boot_params; %ebp, %edi and %ebx must be zero. 1362 136364-bit Boot Protocol 1364==================== 1365 1366For machine with 64bit cpus and 64bit kernel, we could use 64bit bootloader 1367and we need a 64-bit boot protocol. 1368 1369In 64-bit boot protocol, the first step in loading a Linux kernel 1370should be to setup the boot parameters (struct boot_params, 1371traditionally known as "zero page"). The memory for struct boot_params 1372could be allocated anywhere (even above 4G) and initialized to all zero. 1373Then, the setup header at offset 0x01f1 of kernel image on should be 1374loaded into struct boot_params and examined. The end of setup header 1375can be calculated as follows:: 1376 1377 0x0202 + byte value at offset 0x0201 1378 1379In addition to read/modify/write the setup header of the struct 1380boot_params as that of 16-bit boot protocol, the boot loader should 1381also fill the additional fields of the struct boot_params as described 1382in chapter Documentation/x86/zero-page.rst. 1383 1384After setting up the struct boot_params, the boot loader can load 138564-bit kernel in the same way as that of 16-bit boot protocol, but 1386kernel could be loaded above 4G. 1387 1388In 64-bit boot protocol, the kernel is started by jumping to the 138964-bit kernel entry point, which is the start address of loaded 139064-bit kernel plus 0x200. 1391 1392At entry, the CPU must be in 64-bit mode with paging enabled. 1393The range with setup_header.init_size from start address of loaded 1394kernel and zero page and command line buffer get ident mapping; 1395a GDT must be loaded with the descriptors for selectors 1396__BOOT_CS(0x10) and __BOOT_DS(0x18); both descriptors must be 4G flat 1397segment; __BOOT_CS must have execute/read permission, and __BOOT_DS 1398must have read/write permission; CS must be __BOOT_CS and DS, ES, SS 1399must be __BOOT_DS; interrupt must be disabled; %rsi must hold the base 1400address of the struct boot_params. 1401 1402EFI Handover Protocol (deprecated) 1403================================== 1404 1405This protocol allows boot loaders to defer initialisation to the EFI 1406boot stub. The boot loader is required to load the kernel/initrd(s) 1407from the boot media and jump to the EFI handover protocol entry point 1408which is hdr->handover_offset bytes from the beginning of 1409startup_{32,64}. 1410 1411The boot loader MUST respect the kernel's PE/COFF metadata when it comes 1412to section alignment, the memory footprint of the executable image beyond 1413the size of the file itself, and any other aspect of the PE/COFF header 1414that may affect correct operation of the image as a PE/COFF binary in the 1415execution context provided by the EFI firmware. 1416 1417The 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 1422firmware, 'table' is the EFI system table - these are the first two 1423arguments of the "handoff state" as described in section 2.3 of the 1424UEFI specification. 'bp' is the boot loader-allocated boot params. 1425 1426The 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 1432All other fields should be zero. 1433 1434NOTE: 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 1443