1 THE LINUX/x86 BOOT PROTOCOL
2 ---------------------------
4 On the x86 platform, the Linux kernel uses a rather complicated boot
5 convention. This has evolved partially due to historical aspects, as
6 well as the desire in the early days to have the kernel itself be a
7 bootable image, the complicated PC memory model and due to changed
8 expectations in the PC industry caused by the effective demise of
9 real-mode DOS as a mainstream operating system.
11 Currently, the following versions of the Linux/x86 boot protocol exist.
13 Old kernels: zImage/Image support only. Some very early kernels
14 may not even support a command line.
16 Protocol 2.00: (Kernel 1.3.73) Added bzImage and initrd support, as
17 well as a formalized way to communicate between the
18 boot loader and the kernel. setup.S made relocatable,
19 although the traditional setup area still assumed
22 Protocol 2.01: (Kernel 1.3.76) Added a heap overrun warning.
24 Protocol 2.02: (Kernel 2.4.0-test3-pre3) New command line protocol.
25 Lower the conventional memory ceiling. No overwrite
26 of the traditional setup area, thus making booting
27 safe for systems which use the EBDA from SMM or 32-bit
28 BIOS entry points. zImage deprecated but still
31 Protocol 2.03: (Kernel 2.4.18-pre1) Explicitly makes the highest possible
32 initrd address available to the bootloader.
34 Protocol 2.04: (Kernel 2.6.14) Extend the syssize field to four bytes.
36 Protocol 2.05: (Kernel 2.6.20) Make protected mode kernel relocatable.
37 Introduce relocatable_kernel and kernel_alignment fields.
39 Protocol 2.06: (Kernel 2.6.22) Added a field that contains the size of
40 the boot command line.
42 Protocol 2.07: (Kernel 2.6.24) Added paravirtualised boot protocol.
43 Introduced hardware_subarch and hardware_subarch_data
44 and KEEP_SEGMENTS flag in load_flags.
46 Protocol 2.08: (Kernel 2.6.26) Added crc32 checksum and ELF format
47 payload. Introduced payload_offset and payload_length
48 fields to aid in locating the payload.
50 Protocol 2.09: (Kernel 2.6.26) Added a field of 64-bit physical
51 pointer to single linked list of struct setup_data.
53 Protocol 2.10: (Kernel 2.6.31) Added a protocol for relaxed alignment
54 beyond the kernel_alignment added, new init_size and
55 pref_address fields. Added extended boot loader IDs.
57 Protocol 2.11: (Kernel 3.6) Added a field for offset of EFI handover
60 Protocol 2.12: (Kernel 3.9) Added the xloadflags field and extension fields
61 to struct boot_params for for loading bzImage and ramdisk
66 The traditional memory map for the kernel loader, used for Image or
67 zImage kernels, typically looks like:
70 0A0000 +------------------------+
71 | Reserved for BIOS | Do not use. Reserved for BIOS EBDA.
72 09A000 +------------------------+
74 | Stack/heap | For use by the kernel real-mode code.
75 098000 +------------------------+
76 | Kernel setup | The kernel real-mode code.
77 090200 +------------------------+
78 | Kernel boot sector | The kernel legacy boot sector.
79 090000 +------------------------+
80 | Protected-mode kernel | The bulk of the kernel image.
81 010000 +------------------------+
82 | Boot loader | <- Boot sector entry point 0000:7C00
83 001000 +------------------------+
84 | Reserved for MBR/BIOS |
85 000800 +------------------------+
86 | Typically used by MBR |
87 000600 +------------------------+
89 000000 +------------------------+
92 When using bzImage, the protected-mode kernel was relocated to
93 0x100000 ("high memory"), and the kernel real-mode block (boot sector,
94 setup, and stack/heap) was made relocatable to any address between
95 0x10000 and end of low memory. Unfortunately, in protocols 2.00 and
96 2.01 the 0x90000+ memory range is still used internally by the kernel;
97 the 2.02 protocol resolves that problem.
99 It is desirable to keep the "memory ceiling" -- the highest point in
100 low memory touched by the boot loader -- as low as possible, since
101 some newer BIOSes have begun to allocate some rather large amounts of
102 memory, called the Extended BIOS Data Area, near the top of low
103 memory. The boot loader should use the "INT 12h" BIOS call to verify
104 how much low memory is available.
106 Unfortunately, if INT 12h reports that the amount of memory is too
107 low, there is usually nothing the boot loader can do but to report an
108 error to the user. The boot loader should therefore be designed to
109 take up as little space in low memory as it reasonably can. For
110 zImage or old bzImage kernels, which need data written into the
111 0x90000 segment, the boot loader should make sure not to use memory
112 above the 0x9A000 point; too many BIOSes will break above that point.
114 For a modern bzImage kernel with boot protocol version >= 2.02, a
115 memory layout like the following is suggested:
118 | Protected-mode kernel |
119 100000 +------------------------+
121 0A0000 +------------------------+
122 | Reserved for BIOS | Leave as much as possible unused
124 | Command line | (Can also be below the X+10000 mark)
125 X+10000 +------------------------+
126 | Stack/heap | For use by the kernel real-mode code.
127 X+08000 +------------------------+
128 | Kernel setup | The kernel real-mode code.
129 | Kernel boot sector | The kernel legacy boot sector.
130 X +------------------------+
131 | Boot loader | <- Boot sector entry point 0000:7C00
132 001000 +------------------------+
133 | Reserved for MBR/BIOS |
134 000800 +------------------------+
135 | Typically used by MBR |
136 000600 +------------------------+
138 000000 +------------------------+
140 ... where the address X is as low as the design of the boot loader
144 **** THE REAL-MODE KERNEL HEADER
146 In the following text, and anywhere in the kernel boot sequence, "a
147 sector" refers to 512 bytes. It is independent of the actual sector
148 size of the underlying medium.
150 The first step in loading a Linux kernel should be to load the
151 real-mode code (boot sector and setup code) and then examine the
152 following header at offset 0x01f1. The real-mode code can total up to
153 32K, although the boot loader may choose to load only the first two
154 sectors (1K) and then examine the bootup sector size.
156 The header looks like:
158 Offset Proto Name Meaning
161 01F1/1 ALL(1 setup_sects The size of the setup in sectors
162 01F2/2 ALL root_flags If set, the root is mounted readonly
163 01F4/4 2.04+(2 syssize The size of the 32-bit code in 16-byte paras
164 01F8/2 ALL ram_size DO NOT USE - for bootsect.S use only
165 01FA/2 ALL vid_mode Video mode control
166 01FC/2 ALL root_dev Default root device number
167 01FE/2 ALL boot_flag 0xAA55 magic number
168 0200/2 2.00+ jump Jump instruction
169 0202/4 2.00+ header Magic signature "HdrS"
170 0206/2 2.00+ version Boot protocol version supported
171 0208/4 2.00+ realmode_swtch Boot loader hook (see below)
172 020C/2 2.00+ start_sys_seg The load-low segment (0x1000) (obsolete)
173 020E/2 2.00+ kernel_version Pointer to kernel version string
174 0210/1 2.00+ type_of_loader Boot loader identifier
175 0211/1 2.00+ loadflags Boot protocol option flags
176 0212/2 2.00+ setup_move_size Move to high memory size (used with hooks)
177 0214/4 2.00+ code32_start Boot loader hook (see below)
178 0218/4 2.00+ ramdisk_image initrd load address (set by boot loader)
179 021C/4 2.00+ ramdisk_size initrd size (set by boot loader)
180 0220/4 2.00+ bootsect_kludge DO NOT USE - for bootsect.S use only
181 0224/2 2.01+ heap_end_ptr Free memory after setup end
182 0226/1 2.02+(3 ext_loader_ver Extended boot loader version
183 0227/1 2.02+(3 ext_loader_type Extended boot loader ID
184 0228/4 2.02+ cmd_line_ptr 32-bit pointer to the kernel command line
185 022C/4 2.03+ ramdisk_max Highest legal initrd address
186 0230/4 2.05+ kernel_alignment Physical addr alignment required for kernel
187 0234/1 2.05+ relocatable_kernel Whether kernel is relocatable or not
188 0235/1 2.10+ min_alignment Minimum alignment, as a power of two
189 0236/2 2.12+ xloadflags Boot protocol option flags
190 0238/4 2.06+ cmdline_size Maximum size of the kernel command line
191 023C/4 2.07+ hardware_subarch Hardware subarchitecture
192 0240/8 2.07+ hardware_subarch_data Subarchitecture-specific data
193 0248/4 2.08+ payload_offset Offset of kernel payload
194 024C/4 2.08+ payload_length Length of kernel payload
195 0250/8 2.09+ setup_data 64-bit physical pointer to linked list
197 0258/8 2.10+ pref_address Preferred loading address
198 0260/4 2.10+ init_size Linear memory required during initialization
199 0264/4 2.11+ handover_offset Offset of handover entry point
201 (1) For backwards compatibility, if the setup_sects field contains 0, the
204 (2) For boot protocol prior to 2.04, the upper two bytes of the syssize
205 field are unusable, which means the size of a bzImage kernel
206 cannot be determined.
208 (3) Ignored, but safe to set, for boot protocols 2.02-2.09.
210 If the "HdrS" (0x53726448) magic number is not found at offset 0x202,
211 the boot protocol version is "old". Loading an old kernel, the
212 following parameters should be assumed:
216 Real-mode kernel must be located at 0x90000.
218 Otherwise, the "version" field contains the protocol version,
219 e.g. protocol version 2.01 will contain 0x0201 in this field. When
220 setting fields in the header, you must make sure only to set fields
221 supported by the protocol version in use.
224 **** DETAILS OF HEADER FIELDS
226 For each field, some are information from the kernel to the bootloader
227 ("read"), some are expected to be filled out by the bootloader
228 ("write"), and some are expected to be read and modified by the
229 bootloader ("modify").
231 All general purpose boot loaders should write the fields marked
232 (obligatory). Boot loaders who want to load the kernel at a
233 nonstandard address should fill in the fields marked (reloc); other
234 boot loaders can ignore those fields.
236 The byte order of all fields is littleendian (this is x86, after all.)
238 Field name: setup_sects
243 The size of the setup code in 512-byte sectors. If this field is
244 0, the real value is 4. The real-mode code consists of the boot
245 sector (always one 512-byte sector) plus the setup code.
247 Field name: root_flags
248 Type: modify (optional)
252 If this field is nonzero, the root defaults to readonly. The use of
253 this field is deprecated; use the "ro" or "rw" options on the
254 command line instead.
258 Offset/size: 0x1f4/4 (protocol 2.04+) 0x1f4/2 (protocol ALL)
261 The size of the protected-mode code in units of 16-byte paragraphs.
262 For protocol versions older than 2.04 this field is only two bytes
263 wide, and therefore cannot be trusted for the size of a kernel if
264 the LOAD_HIGH flag is set.
267 Type: kernel internal
271 This field is obsolete.
274 Type: modify (obligatory)
277 Please see the section on SPECIAL COMMAND LINE OPTIONS.
280 Type: modify (optional)
284 The default root device device number. The use of this field is
285 deprecated, use the "root=" option on the command line instead.
287 Field name: boot_flag
292 Contains 0xAA55. This is the closest thing old Linux kernels have
300 Contains an x86 jump instruction, 0xEB followed by a signed offset
301 relative to byte 0x202. This can be used to determine the size of
309 Contains the magic number "HdrS" (0x53726448).
316 Contains the boot protocol version, in (major << 8)+minor format,
317 e.g. 0x0204 for version 2.04, and 0x0a11 for a hypothetical version
320 Field name: realmode_swtch
321 Type: modify (optional)
325 Boot loader hook (see ADVANCED BOOT LOADER HOOKS below.)
327 Field name: start_sys_seg
332 The load low segment (0x1000). Obsolete.
334 Field name: kernel_version
339 If set to a nonzero value, contains a pointer to a NUL-terminated
340 human-readable kernel version number string, less 0x200. This can
341 be used to display the kernel version to the user. This value
342 should be less than (0x200*setup_sects).
344 For example, if this value is set to 0x1c00, the kernel version
345 number string can be found at offset 0x1e00 in the kernel file.
346 This is a valid value if and only if the "setup_sects" field
347 contains the value 15 or higher, as:
349 0x1c00 < 15*0x200 (= 0x1e00) but
350 0x1c00 >= 14*0x200 (= 0x1c00)
352 0x1c00 >> 9 = 14, so the minimum value for setup_secs is 15.
354 Field name: type_of_loader
355 Type: write (obligatory)
359 If your boot loader has an assigned id (see table below), enter
360 0xTV here, where T is an identifier for the boot loader and V is
361 a version number. Otherwise, enter 0xFF here.
363 For boot loader IDs above T = 0xD, write T = 0xE to this field and
364 write the extended ID minus 0x10 to the ext_loader_type field.
365 Similarly, the ext_loader_ver field can be used to provide more than
366 four bits for the bootloader version.
368 For example, for T = 0x15, V = 0x234, write:
370 type_of_loader <- 0xE4
371 ext_loader_type <- 0x05
372 ext_loader_ver <- 0x23
374 Assigned boot loader ids (hexadecimal):
376 0 LILO (0x00 reserved for pre-2.00 bootloader)
378 2 bootsect-loader (0x20, all other values reserved)
380 4 Etherboot/gPXE/iPXE
387 C Arcturus Networks uCbootloader
389 E Extended (see ext_loader_type)
390 F Special (0xFF = undefined)
392 11 Minimal Linux Bootloader <http://sebastian-plotz.blogspot.de>
394 Please contact <hpa@zytor.com> if you need a bootloader ID
397 Field name: loadflags
398 Type: modify (obligatory)
402 This field is a bitmask.
404 Bit 0 (read): LOADED_HIGH
405 - If 0, the protected-mode code is loaded at 0x10000.
406 - If 1, the protected-mode code is loaded at 0x100000.
408 Bit 5 (write): QUIET_FLAG
409 - If 0, print early messages.
410 - If 1, suppress early messages.
411 This requests to the kernel (decompressor and early
412 kernel) to not write early messages that require
413 accessing the display hardware directly.
415 Bit 6 (write): KEEP_SEGMENTS
417 - If 0, reload the segment registers in the 32bit entry point.
418 - If 1, do not reload the segment registers in the 32bit entry point.
419 Assume that %cs %ds %ss %es are all set to flat segments with
420 a base of 0 (or the equivalent for their environment).
422 Bit 7 (write): CAN_USE_HEAP
423 Set this bit to 1 to indicate that the value entered in the
424 heap_end_ptr is valid. If this field is clear, some setup code
425 functionality will be disabled.
427 Field name: setup_move_size
428 Type: modify (obligatory)
432 When using protocol 2.00 or 2.01, if the real mode kernel is not
433 loaded at 0x90000, it gets moved there later in the loading
434 sequence. Fill in this field if you want additional data (such as
435 the kernel command line) moved in addition to the real-mode kernel
438 The unit is bytes starting with the beginning of the boot sector.
440 This field is can be ignored when the protocol is 2.02 or higher, or
441 if the real-mode code is loaded at 0x90000.
443 Field name: code32_start
444 Type: modify (optional, reloc)
448 The address to jump to in protected mode. This defaults to the load
449 address of the kernel, and can be used by the boot loader to
450 determine the proper load address.
452 This field can be modified for two purposes:
454 1. as a boot loader hook (see ADVANCED BOOT LOADER HOOKS below.)
456 2. if a bootloader which does not install a hook loads a
457 relocatable kernel at a nonstandard address it will have to modify
458 this field to point to the load address.
460 Field name: ramdisk_image
461 Type: write (obligatory)
465 The 32-bit linear address of the initial ramdisk or ramfs. Leave at
466 zero if there is no initial ramdisk/ramfs.
468 Field name: ramdisk_size
469 Type: write (obligatory)
473 Size of the initial ramdisk or ramfs. Leave at zero if there is no
474 initial ramdisk/ramfs.
476 Field name: bootsect_kludge
477 Type: kernel internal
481 This field is obsolete.
483 Field name: heap_end_ptr
484 Type: write (obligatory)
488 Set this field to the offset (from the beginning of the real-mode
489 code) of the end of the setup stack/heap, minus 0x0200.
491 Field name: ext_loader_ver
492 Type: write (optional)
496 This field is used as an extension of the version number in the
497 type_of_loader field. The total version number is considered to be
498 (type_of_loader & 0x0f) + (ext_loader_ver << 4).
500 The use of this field is boot loader specific. If not written, it
503 Kernels prior to 2.6.31 did not recognize this field, but it is safe
504 to write for protocol version 2.02 or higher.
506 Field name: ext_loader_type
507 Type: write (obligatory if (type_of_loader & 0xf0) == 0xe0)
511 This field is used as an extension of the type number in
512 type_of_loader field. If the type in type_of_loader is 0xE, then
513 the actual type is (ext_loader_type + 0x10).
515 This field is ignored if the type in type_of_loader is not 0xE.
517 Kernels prior to 2.6.31 did not recognize this field, but it is safe
518 to write for protocol version 2.02 or higher.
520 Field name: cmd_line_ptr
521 Type: write (obligatory)
525 Set this field to the linear address of the kernel command line.
526 The kernel command line can be located anywhere between the end of
527 the setup heap and 0xA0000; it does not have to be located in the
528 same 64K segment as the real-mode code itself.
530 Fill in this field even if your boot loader does not support a
531 command line, in which case you can point this to an empty string
532 (or better yet, to the string "auto".) If this field is left at
533 zero, the kernel will assume that your boot loader does not support
536 Field name: ramdisk_max
541 The maximum address that may be occupied by the initial
542 ramdisk/ramfs contents. For boot protocols 2.02 or earlier, this
543 field is not present, and the maximum address is 0x37FFFFFF. (This
544 address is defined as the address of the highest safe byte, so if
545 your ramdisk is exactly 131072 bytes long and this field is
546 0x37FFFFFF, you can start your ramdisk at 0x37FE0000.)
548 Field name: kernel_alignment
549 Type: read/modify (reloc)
551 Protocol: 2.05+ (read), 2.10+ (modify)
553 Alignment unit required by the kernel (if relocatable_kernel is
554 true.) A relocatable kernel that is loaded at an alignment
555 incompatible with the value in this field will be realigned during
556 kernel initialization.
558 Starting with protocol version 2.10, this reflects the kernel
559 alignment preferred for optimal performance; it is possible for the
560 loader to modify this field to permit a lesser alignment. See the
561 min_alignment and pref_address field below.
563 Field name: relocatable_kernel
568 If this field is nonzero, the protected-mode part of the kernel can
569 be loaded at any address that satisfies the kernel_alignment field.
570 After loading, the boot loader must set the code32_start field to
571 point to the loaded code, or to a boot loader hook.
573 Field name: min_alignment
578 This field, if nonzero, indicates as a power of two the minimum
579 alignment required, as opposed to preferred, by the kernel to boot.
580 If a boot loader makes use of this field, it should update the
581 kernel_alignment field with the alignment unit desired; typically:
583 kernel_alignment = 1 << min_alignment
585 There may be a considerable performance cost with an excessively
586 misaligned kernel. Therefore, a loader should typically try each
587 power-of-two alignment from kernel_alignment down to this alignment.
589 Field name: xloadflags
594 This field is a bitmask.
596 Bit 0 (read): XLF_KERNEL_64
597 - If 1, this kernel has the legacy 64-bit entry point at 0x200.
599 Bit 1 (read): XLF_CAN_BE_LOADED_ABOVE_4G
600 - If 1, kernel/boot_params/cmdline/ramdisk can be above 4G.
602 Bit 2 (read): XLF_EFI_HANDOVER_32
603 - If 1, the kernel supports the 32-bit EFI handoff entry point
604 given at handover_offset.
606 Bit 3 (read): XLF_EFI_HANDOVER_64
607 - If 1, the kernel supports the 64-bit EFI handoff entry point
608 given at handover_offset + 0x200.
610 Field name: cmdline_size
615 The maximum size of the command line without the terminating
616 zero. This means that the command line can contain at most
617 cmdline_size characters. With protocol version 2.05 and earlier, the
618 maximum size was 255.
620 Field name: hardware_subarch
621 Type: write (optional, defaults to x86/PC)
625 In a paravirtualized environment the hardware low level architectural
626 pieces such as interrupt handling, page table handling, and
627 accessing process control registers needs to be done differently.
629 This field allows the bootloader to inform the kernel we are in one
630 one of those environments.
632 0x00000000 The default x86/PC environment
635 0x00000003 Moorestown MID
636 0x00000004 CE4100 TV Platform
638 Field name: hardware_subarch_data
639 Type: write (subarch-dependent)
643 A pointer to data that is specific to hardware subarch
644 This field is currently unused for the default x86/PC environment,
647 Field name: payload_offset
652 If non-zero then this field contains the offset from the beginning
653 of the protected-mode code to the payload.
655 The payload may be compressed. The format of both the compressed and
656 uncompressed data should be determined using the standard magic
657 numbers. The currently supported compression formats are gzip
658 (magic numbers 1F 8B or 1F 9E), bzip2 (magic number 42 5A), LZMA
659 (magic number 5D 00), and XZ (magic number FD 37). The uncompressed
660 payload is currently always ELF (magic number 7F 45 4C 46).
662 Field name: payload_length
667 The length of the payload.
669 Field name: setup_data
670 Type: write (special)
674 The 64-bit physical pointer to NULL terminated single linked list of
675 struct setup_data. This is used to define a more extensible boot
676 parameters passing mechanism. The definition of struct setup_data is
686 Where, the next is a 64-bit physical pointer to the next node of
687 linked list, the next field of the last node is 0; the type is used
688 to identify the contents of data; the len is the length of data
689 field; the data holds the real payload.
691 This list may be modified at a number of points during the bootup
692 process. Therefore, when modifying this list one should always make
693 sure to consider the case where the linked list already contains
696 Field name: pref_address
701 This field, if nonzero, represents a preferred load address for the
702 kernel. A relocating bootloader should attempt to load at this
705 A non-relocatable kernel will unconditionally move itself and to run
708 Field name: init_size
712 This field indicates the amount of linear contiguous memory starting
713 at the kernel runtime start address that the kernel needs before it
714 is capable of examining its memory map. This is not the same thing
715 as the total amount of memory the kernel needs to boot, but it can
716 be used by a relocating boot loader to help select a safe load
717 address for the kernel.
719 The kernel runtime start address is determined by the following algorithm:
721 if (relocatable_kernel)
722 runtime_start = align_up(load_address, kernel_alignment)
724 runtime_start = pref_address
726 Field name: handover_offset
730 This field is the offset from the beginning of the kernel image to
731 the EFI handover protocol entry point. Boot loaders using the EFI
732 handover protocol to boot the kernel should jump to this offset.
734 See EFI HANDOVER PROTOCOL below for more details.
737 **** THE IMAGE CHECKSUM
739 From boot protocol version 2.08 onwards the CRC-32 is calculated over
740 the entire file using the characteristic polynomial 0x04C11DB7 and an
741 initial remainder of 0xffffffff. The checksum is appended to the
742 file; therefore the CRC of the file up to the limit specified in the
743 syssize field of the header is always 0.
746 **** THE KERNEL COMMAND LINE
748 The kernel command line has become an important way for the boot
749 loader to communicate with the kernel. Some of its options are also
750 relevant to the boot loader itself, see "special command line options"
753 The kernel command line is a null-terminated string. The maximum
754 length can be retrieved from the field cmdline_size. Before protocol
755 version 2.06, the maximum was 255 characters. A string that is too
756 long will be automatically truncated by the kernel.
758 If the boot protocol version is 2.02 or later, the address of the
759 kernel command line is given by the header field cmd_line_ptr (see
760 above.) This address can be anywhere between the end of the setup
763 If the protocol version is *not* 2.02 or higher, the kernel
764 command line is entered using the following protocol:
766 At offset 0x0020 (word), "cmd_line_magic", enter the magic
769 At offset 0x0022 (word), "cmd_line_offset", enter the offset
770 of the kernel command line (relative to the start of the
773 The kernel command line *must* be within the memory region
774 covered by setup_move_size, so you may need to adjust this
778 **** MEMORY LAYOUT OF THE REAL-MODE CODE
780 The real-mode code requires a stack/heap to be set up, as well as
781 memory allocated for the kernel command line. This needs to be done
782 in the real-mode accessible memory in bottom megabyte.
784 It should be noted that modern machines often have a sizable Extended
785 BIOS Data Area (EBDA). As a result, it is advisable to use as little
786 of the low megabyte as possible.
788 Unfortunately, under the following circumstances the 0x90000 memory
789 segment has to be used:
791 - When loading a zImage kernel ((loadflags & 0x01) == 0).
792 - When loading a 2.01 or earlier boot protocol kernel.
794 -> For the 2.00 and 2.01 boot protocols, the real-mode code
795 can be loaded at another address, but it is internally
796 relocated to 0x90000. For the "old" protocol, the
797 real-mode code must be loaded at 0x90000.
799 When loading at 0x90000, avoid using memory above 0x9a000.
801 For boot protocol 2.02 or higher, the command line does not have to be
802 located in the same 64K segment as the real-mode setup code; it is
803 thus permitted to give the stack/heap the full 64K segment and locate
804 the command line above it.
806 The kernel command line should not be located below the real-mode
807 code, nor should it be located in high memory.
810 **** SAMPLE BOOT CONFIGURATION
812 As a sample configuration, assume the following layout of the real
815 When loading below 0x90000, use the entire segment:
817 0x0000-0x7fff Real mode kernel
818 0x8000-0xdfff Stack and heap
819 0xe000-0xffff Kernel command line
821 When loading at 0x90000 OR the protocol version is 2.01 or earlier:
823 0x0000-0x7fff Real mode kernel
824 0x8000-0x97ff Stack and heap
825 0x9800-0x9fff Kernel command line
827 Such a boot loader should enter the following fields in the header:
829 unsigned long base_ptr; /* base address for real-mode segment */
831 if ( setup_sects == 0 ) {
835 if ( protocol >= 0x0200 ) {
836 type_of_loader = <type code>;
837 if ( loading_initrd ) {
838 ramdisk_image = <initrd_address>;
839 ramdisk_size = <initrd_size>;
842 if ( protocol >= 0x0202 && loadflags & 0x01 )
847 if ( protocol >= 0x0201 ) {
848 heap_end_ptr = heap_end - 0x200;
849 loadflags |= 0x80; /* CAN_USE_HEAP */
852 if ( protocol >= 0x0202 ) {
853 cmd_line_ptr = base_ptr + heap_end;
854 strcpy(cmd_line_ptr, cmdline);
856 cmd_line_magic = 0xA33F;
857 cmd_line_offset = heap_end;
858 setup_move_size = heap_end + strlen(cmdline)+1;
859 strcpy(base_ptr+cmd_line_offset, cmdline);
862 /* Very old kernel */
866 cmd_line_magic = 0xA33F;
867 cmd_line_offset = heap_end;
869 /* A very old kernel MUST have its real-mode code
872 if ( base_ptr != 0x90000 ) {
873 /* Copy the real-mode kernel */
874 memcpy(0x90000, base_ptr, (setup_sects+1)*512);
875 base_ptr = 0x90000; /* Relocated */
878 strcpy(0x90000+cmd_line_offset, cmdline);
880 /* It is recommended to clear memory up to the 32K mark */
881 memset(0x90000 + (setup_sects+1)*512, 0,
882 (64-(setup_sects+1))*512);
886 **** LOADING THE REST OF THE KERNEL
888 The 32-bit (non-real-mode) kernel starts at offset (setup_sects+1)*512
889 in the kernel file (again, if setup_sects == 0 the real value is 4.)
890 It should be loaded at address 0x10000 for Image/zImage kernels and
891 0x100000 for bzImage kernels.
893 The kernel is a bzImage kernel if the protocol >= 2.00 and the 0x01
894 bit (LOAD_HIGH) in the loadflags field is set:
896 is_bzImage = (protocol >= 0x0200) && (loadflags & 0x01);
897 load_address = is_bzImage ? 0x100000 : 0x10000;
899 Note that Image/zImage kernels can be up to 512K in size, and thus use
900 the entire 0x10000-0x90000 range of memory. This means it is pretty
901 much a requirement for these kernels to load the real-mode part at
902 0x90000. bzImage kernels allow much more flexibility.
905 **** SPECIAL COMMAND LINE OPTIONS
907 If the command line provided by the boot loader is entered by the
908 user, the user may expect the following command line options to work.
909 They should normally not be deleted from the kernel command line even
910 though not all of them are actually meaningful to the kernel. Boot
911 loader authors who need additional command line options for the boot
912 loader itself should get them registered in
913 Documentation/kernel-parameters.txt to make sure they will not
914 conflict with actual kernel options now or in the future.
917 <mode> here is either an integer (in C notation, either
918 decimal, octal, or hexadecimal) or one of the strings
919 "normal" (meaning 0xFFFF), "ext" (meaning 0xFFFE) or "ask"
920 (meaning 0xFFFD). This value should be entered into the
921 vid_mode field, as it is used by the kernel before the command
925 <size> is an integer in C notation optionally followed by
926 (case insensitive) K, M, G, T, P or E (meaning << 10, << 20,
927 << 30, << 40, << 50 or << 60). This specifies the end of
928 memory to the kernel. This affects the possible placement of
929 an initrd, since an initrd should be placed near end of
930 memory. Note that this is an option to *both* the kernel and
934 An initrd should be loaded. The meaning of <file> is
935 obviously bootloader-dependent, and some boot loaders
936 (e.g. LILO) do not have such a command.
938 In addition, some boot loaders add the following options to the
939 user-specified command line:
942 The boot image which was loaded. Again, the meaning of <file>
943 is obviously bootloader-dependent.
946 The kernel was booted without explicit user intervention.
948 If these options are added by the boot loader, it is highly
949 recommended that they are located *first*, before the user-specified
950 or configuration-specified command line. Otherwise, "init=/bin/sh"
951 gets confused by the "auto" option.
954 **** RUNNING THE KERNEL
956 The kernel is started by jumping to the kernel entry point, which is
957 located at *segment* offset 0x20 from the start of the real mode
958 kernel. This means that if you loaded your real-mode kernel code at
959 0x90000, the kernel entry point is 9020:0000.
961 At entry, ds = es = ss should point to the start of the real-mode
962 kernel code (0x9000 if the code is loaded at 0x90000), sp should be
963 set up properly, normally pointing to the top of the heap, and
964 interrupts should be disabled. Furthermore, to guard against bugs in
965 the kernel, it is recommended that the boot loader sets fs = gs = ds =
968 In our example from above, we would do:
970 /* Note: in the case of the "old" kernel protocol, base_ptr must
971 be == 0x90000 at this point; see the previous sample code */
975 cli(); /* Enter with interrupts disabled! */
977 /* Set up the real-mode kernel stack */
981 _DS = _ES = _FS = _GS = seg;
982 jmp_far(seg+0x20, 0); /* Run the kernel */
984 If your boot sector accesses a floppy drive, it is recommended to
985 switch off the floppy motor before running the kernel, since the
986 kernel boot leaves interrupts off and thus the motor will not be
987 switched off, especially if the loaded kernel has the floppy driver as
988 a demand-loaded module!
991 **** ADVANCED BOOT LOADER HOOKS
993 If the boot loader runs in a particularly hostile environment (such as
994 LOADLIN, which runs under DOS) it may be impossible to follow the
995 standard memory location requirements. Such a boot loader may use the
996 following hooks that, if set, are invoked by the kernel at the
997 appropriate time. The use of these hooks should probably be
998 considered an absolutely last resort!
1000 IMPORTANT: All the hooks are required to preserve %esp, %ebp, %esi and
1001 %edi across invocation.
1004 A 16-bit real mode far subroutine invoked immediately before
1005 entering protected mode. The default routine disables NMI, so
1006 your routine should probably do so, too.
1009 A 32-bit flat-mode routine *jumped* to immediately after the
1010 transition to protected mode, but before the kernel is
1011 uncompressed. No segments, except CS, are guaranteed to be
1012 set up (current kernels do, but older ones do not); you should
1013 set them up to BOOT_DS (0x18) yourself.
1015 After completing your hook, you should jump to the address
1016 that was in this field before your boot loader overwrote it
1017 (relocated, if appropriate.)
1020 **** 32-bit BOOT PROTOCOL
1022 For machine with some new BIOS other than legacy BIOS, such as EFI,
1023 LinuxBIOS, etc, and kexec, the 16-bit real mode setup code in kernel
1024 based on legacy BIOS can not be used, so a 32-bit boot protocol needs
1027 In 32-bit boot protocol, the first step in loading a Linux kernel
1028 should be to setup the boot parameters (struct boot_params,
1029 traditionally known as "zero page"). The memory for struct boot_params
1030 should be allocated and initialized to all zero. Then the setup header
1031 from offset 0x01f1 of kernel image on should be loaded into struct
1032 boot_params and examined. The end of setup header can be calculated as
1035 0x0202 + byte value at offset 0x0201
1037 In addition to read/modify/write the setup header of the struct
1038 boot_params as that of 16-bit boot protocol, the boot loader should
1039 also fill the additional fields of the struct boot_params as that
1040 described in zero-page.txt.
1042 After setting up the struct boot_params, the boot loader can load the
1043 32/64-bit kernel in the same way as that of 16-bit boot protocol.
1045 In 32-bit boot protocol, the kernel is started by jumping to the
1046 32-bit kernel entry point, which is the start address of loaded
1049 At entry, the CPU must be in 32-bit protected mode with paging
1050 disabled; a GDT must be loaded with the descriptors for selectors
1051 __BOOT_CS(0x10) and __BOOT_DS(0x18); both descriptors must be 4G flat
1052 segment; __BOOT_CS must have execute/read permission, and __BOOT_DS
1053 must have read/write permission; CS must be __BOOT_CS and DS, ES, SS
1054 must be __BOOT_DS; interrupt must be disabled; %esi must hold the base
1055 address of the struct boot_params; %ebp, %edi and %ebx must be zero.
1057 **** EFI HANDOVER PROTOCOL
1059 This protocol allows boot loaders to defer initialisation to the EFI
1060 boot stub. The boot loader is required to load the kernel/initrd(s)
1061 from the boot media and jump to the EFI handover protocol entry point
1062 which is hdr->handover_offset bytes from the beginning of
1065 The function prototype for the handover entry point looks like this,
1067 efi_main(void *handle, efi_system_table_t *table, struct boot_params *bp)
1069 'handle' is the EFI image handle passed to the boot loader by the EFI
1070 firmware, 'table' is the EFI system table - these are the first two
1071 arguments of the "handoff state" as described in section 2.3 of the
1072 UEFI specification. 'bp' is the boot loader-allocated boot params.
1074 The boot loader *must* fill out the following fields in bp,
1079 o hdr.ramdisk_image (if applicable)
1080 o hdr.ramdisk_size (if applicable)
1082 All other fields should be zero.