boot.txt 34 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013
  1. THE LINUX/x86 BOOT PROTOCOL
  2. ---------------------------
  3. On the x86 platform, the Linux kernel uses a rather complicated boot
  4. convention. This has evolved partially due to historical aspects, as
  5. well as the desire in the early days to have the kernel itself be a
  6. bootable image, the complicated PC memory model and due to changed
  7. expectations in the PC industry caused by the effective demise of
  8. real-mode DOS as a mainstream operating system.
  9. Currently, the following versions of the Linux/x86 boot protocol exist.
  10. Old kernels: zImage/Image support only. Some very early kernels
  11. may not even support a command line.
  12. Protocol 2.00: (Kernel 1.3.73) Added bzImage and initrd support, as
  13. well as a formalized way to communicate between the
  14. boot loader and the kernel. setup.S made relocatable,
  15. although the traditional setup area still assumed
  16. writable.
  17. Protocol 2.01: (Kernel 1.3.76) Added a heap overrun warning.
  18. Protocol 2.02: (Kernel 2.4.0-test3-pre3) New command line protocol.
  19. Lower the conventional memory ceiling. No overwrite
  20. of the traditional setup area, thus making booting
  21. safe for systems which use the EBDA from SMM or 32-bit
  22. BIOS entry points. zImage deprecated but still
  23. supported.
  24. Protocol 2.03: (Kernel 2.4.18-pre1) Explicitly makes the highest possible
  25. initrd address available to the bootloader.
  26. Protocol 2.04: (Kernel 2.6.14) Extend the syssize field to four bytes.
  27. Protocol 2.05: (Kernel 2.6.20) Make protected mode kernel relocatable.
  28. Introduce relocatable_kernel and kernel_alignment fields.
  29. Protocol 2.06: (Kernel 2.6.22) Added a field that contains the size of
  30. the boot command line.
  31. Protocol 2.07: (Kernel 2.6.24) Added paravirtualised boot protocol.
  32. Introduced hardware_subarch and hardware_subarch_data
  33. and KEEP_SEGMENTS flag in load_flags.
  34. Protocol 2.08: (Kernel 2.6.26) Added crc32 checksum and ELF format
  35. payload. Introduced payload_offset and payload_length
  36. fields to aid in locating the payload.
  37. Protocol 2.09: (Kernel 2.6.26) Added a field of 64-bit physical
  38. pointer to single linked list of struct setup_data.
  39. Protocol 2.10: (Kernel 2.6.31) Added a protocol for relaxed alignment
  40. beyond the kernel_alignment added, new init_size and
  41. pref_address fields. Added extended boot loader IDs.
  42. **** MEMORY LAYOUT
  43. The traditional memory map for the kernel loader, used for Image or
  44. zImage kernels, typically looks like:
  45. | |
  46. 0A0000 +------------------------+
  47. | Reserved for BIOS | Do not use. Reserved for BIOS EBDA.
  48. 09A000 +------------------------+
  49. | Command line |
  50. | Stack/heap | For use by the kernel real-mode code.
  51. 098000 +------------------------+
  52. | Kernel setup | The kernel real-mode code.
  53. 090200 +------------------------+
  54. | Kernel boot sector | The kernel legacy boot sector.
  55. 090000 +------------------------+
  56. | Protected-mode kernel | The bulk of the kernel image.
  57. 010000 +------------------------+
  58. | Boot loader | <- Boot sector entry point 0000:7C00
  59. 001000 +------------------------+
  60. | Reserved for MBR/BIOS |
  61. 000800 +------------------------+
  62. | Typically used by MBR |
  63. 000600 +------------------------+
  64. | BIOS use only |
  65. 000000 +------------------------+
  66. When using bzImage, the protected-mode kernel was relocated to
  67. 0x100000 ("high memory"), and the kernel real-mode block (boot sector,
  68. setup, and stack/heap) was made relocatable to any address between
  69. 0x10000 and end of low memory. Unfortunately, in protocols 2.00 and
  70. 2.01 the 0x90000+ memory range is still used internally by the kernel;
  71. the 2.02 protocol resolves that problem.
  72. It is desirable to keep the "memory ceiling" -- the highest point in
  73. low memory touched by the boot loader -- as low as possible, since
  74. some newer BIOSes have begun to allocate some rather large amounts of
  75. memory, called the Extended BIOS Data Area, near the top of low
  76. memory. The boot loader should use the "INT 12h" BIOS call to verify
  77. how much low memory is available.
  78. Unfortunately, if INT 12h reports that the amount of memory is too
  79. low, there is usually nothing the boot loader can do but to report an
  80. error to the user. The boot loader should therefore be designed to
  81. take up as little space in low memory as it reasonably can. For
  82. zImage or old bzImage kernels, which need data written into the
  83. 0x90000 segment, the boot loader should make sure not to use memory
  84. above the 0x9A000 point; too many BIOSes will break above that point.
  85. For a modern bzImage kernel with boot protocol version >= 2.02, a
  86. memory layout like the following is suggested:
  87. ~ ~
  88. | Protected-mode kernel |
  89. 100000 +------------------------+
  90. | I/O memory hole |
  91. 0A0000 +------------------------+
  92. | Reserved for BIOS | Leave as much as possible unused
  93. ~ ~
  94. | Command line | (Can also be below the X+10000 mark)
  95. X+10000 +------------------------+
  96. | Stack/heap | For use by the kernel real-mode code.
  97. X+08000 +------------------------+
  98. | Kernel setup | The kernel real-mode code.
  99. | Kernel boot sector | The kernel legacy boot sector.
  100. X +------------------------+
  101. | Boot loader | <- Boot sector entry point 0000:7C00
  102. 001000 +------------------------+
  103. | Reserved for MBR/BIOS |
  104. 000800 +------------------------+
  105. | Typically used by MBR |
  106. 000600 +------------------------+
  107. | BIOS use only |
  108. 000000 +------------------------+
  109. ... where the address X is as low as the design of the boot loader
  110. permits.
  111. **** THE REAL-MODE KERNEL HEADER
  112. In the following text, and anywhere in the kernel boot sequence, "a
  113. sector" refers to 512 bytes. It is independent of the actual sector
  114. size of the underlying medium.
  115. The first step in loading a Linux kernel should be to load the
  116. real-mode code (boot sector and setup code) and then examine the
  117. following header at offset 0x01f1. The real-mode code can total up to
  118. 32K, although the boot loader may choose to load only the first two
  119. sectors (1K) and then examine the bootup sector size.
  120. The header looks like:
  121. Offset Proto Name Meaning
  122. /Size
  123. 01F1/1 ALL(1 setup_sects The size of the setup in sectors
  124. 01F2/2 ALL root_flags If set, the root is mounted readonly
  125. 01F4/4 2.04+(2 syssize The size of the 32-bit code in 16-byte paras
  126. 01F8/2 ALL ram_size DO NOT USE - for bootsect.S use only
  127. 01FA/2 ALL vid_mode Video mode control
  128. 01FC/2 ALL root_dev Default root device number
  129. 01FE/2 ALL boot_flag 0xAA55 magic number
  130. 0200/2 2.00+ jump Jump instruction
  131. 0202/4 2.00+ header Magic signature "HdrS"
  132. 0206/2 2.00+ version Boot protocol version supported
  133. 0208/4 2.00+ realmode_swtch Boot loader hook (see below)
  134. 020C/2 2.00+ start_sys_seg The load-low segment (0x1000) (obsolete)
  135. 020E/2 2.00+ kernel_version Pointer to kernel version string
  136. 0210/1 2.00+ type_of_loader Boot loader identifier
  137. 0211/1 2.00+ loadflags Boot protocol option flags
  138. 0212/2 2.00+ setup_move_size Move to high memory size (used with hooks)
  139. 0214/4 2.00+ code32_start Boot loader hook (see below)
  140. 0218/4 2.00+ ramdisk_image initrd load address (set by boot loader)
  141. 021C/4 2.00+ ramdisk_size initrd size (set by boot loader)
  142. 0220/4 2.00+ bootsect_kludge DO NOT USE - for bootsect.S use only
  143. 0224/2 2.01+ heap_end_ptr Free memory after setup end
  144. 0226/1 2.02+(3 ext_loader_ver Extended boot loader version
  145. 0227/1 2.02+(3 ext_loader_type Extended boot loader ID
  146. 0228/4 2.02+ cmd_line_ptr 32-bit pointer to the kernel command line
  147. 022C/4 2.03+ ramdisk_max Highest legal initrd address
  148. 0230/4 2.05+ kernel_alignment Physical addr alignment required for kernel
  149. 0234/1 2.05+ relocatable_kernel Whether kernel is relocatable or not
  150. 0235/1 2.10+ min_alignment Minimum alignment, as a power of two
  151. 0236/2 N/A pad3 Unused
  152. 0238/4 2.06+ cmdline_size Maximum size of the kernel command line
  153. 023C/4 2.07+ hardware_subarch Hardware subarchitecture
  154. 0240/8 2.07+ hardware_subarch_data Subarchitecture-specific data
  155. 0248/4 2.08+ payload_offset Offset of kernel payload
  156. 024C/4 2.08+ payload_length Length of kernel payload
  157. 0250/8 2.09+ setup_data 64-bit physical pointer to linked list
  158. of struct setup_data
  159. 0258/8 2.10+ pref_address Preferred loading address
  160. 0260/4 2.10+ init_size Linear memory required during initialization
  161. (1) For backwards compatibility, if the setup_sects field contains 0, the
  162. real value is 4.
  163. (2) For boot protocol prior to 2.04, the upper two bytes of the syssize
  164. field are unusable, which means the size of a bzImage kernel
  165. cannot be determined.
  166. (3) Ignored, but safe to set, for boot protocols 2.02-2.09.
  167. If the "HdrS" (0x53726448) magic number is not found at offset 0x202,
  168. the boot protocol version is "old". Loading an old kernel, the
  169. following parameters should be assumed:
  170. Image type = zImage
  171. initrd not supported
  172. Real-mode kernel must be located at 0x90000.
  173. Otherwise, the "version" field contains the protocol version,
  174. e.g. protocol version 2.01 will contain 0x0201 in this field. When
  175. setting fields in the header, you must make sure only to set fields
  176. supported by the protocol version in use.
  177. **** DETAILS OF HEADER FIELDS
  178. For each field, some are information from the kernel to the bootloader
  179. ("read"), some are expected to be filled out by the bootloader
  180. ("write"), and some are expected to be read and modified by the
  181. bootloader ("modify").
  182. All general purpose boot loaders should write the fields marked
  183. (obligatory). Boot loaders who want to load the kernel at a
  184. nonstandard address should fill in the fields marked (reloc); other
  185. boot loaders can ignore those fields.
  186. The byte order of all fields is littleendian (this is x86, after all.)
  187. Field name: setup_sects
  188. Type: read
  189. Offset/size: 0x1f1/1
  190. Protocol: ALL
  191. The size of the setup code in 512-byte sectors. If this field is
  192. 0, the real value is 4. The real-mode code consists of the boot
  193. sector (always one 512-byte sector) plus the setup code.
  194. Field name: root_flags
  195. Type: modify (optional)
  196. Offset/size: 0x1f2/2
  197. Protocol: ALL
  198. If this field is nonzero, the root defaults to readonly. The use of
  199. this field is deprecated; use the "ro" or "rw" options on the
  200. command line instead.
  201. Field name: syssize
  202. Type: read
  203. Offset/size: 0x1f4/4 (protocol 2.04+) 0x1f4/2 (protocol ALL)
  204. Protocol: 2.04+
  205. The size of the protected-mode code in units of 16-byte paragraphs.
  206. For protocol versions older than 2.04 this field is only two bytes
  207. wide, and therefore cannot be trusted for the size of a kernel if
  208. the LOAD_HIGH flag is set.
  209. Field name: ram_size
  210. Type: kernel internal
  211. Offset/size: 0x1f8/2
  212. Protocol: ALL
  213. This field is obsolete.
  214. Field name: vid_mode
  215. Type: modify (obligatory)
  216. Offset/size: 0x1fa/2
  217. Please see the section on SPECIAL COMMAND LINE OPTIONS.
  218. Field name: root_dev
  219. Type: modify (optional)
  220. Offset/size: 0x1fc/2
  221. Protocol: ALL
  222. The default root device device number. The use of this field is
  223. deprecated, use the "root=" option on the command line instead.
  224. Field name: boot_flag
  225. Type: read
  226. Offset/size: 0x1fe/2
  227. Protocol: ALL
  228. Contains 0xAA55. This is the closest thing old Linux kernels have
  229. to a magic number.
  230. Field name: jump
  231. Type: read
  232. Offset/size: 0x200/2
  233. Protocol: 2.00+
  234. Contains an x86 jump instruction, 0xEB followed by a signed offset
  235. relative to byte 0x202. This can be used to determine the size of
  236. the header.
  237. Field name: header
  238. Type: read
  239. Offset/size: 0x202/4
  240. Protocol: 2.00+
  241. Contains the magic number "HdrS" (0x53726448).
  242. Field name: version
  243. Type: read
  244. Offset/size: 0x206/2
  245. Protocol: 2.00+
  246. Contains the boot protocol version, in (major << 8)+minor format,
  247. e.g. 0x0204 for version 2.04, and 0x0a11 for a hypothetical version
  248. 10.17.
  249. Field name: realmode_swtch
  250. Type: modify (optional)
  251. Offset/size: 0x208/4
  252. Protocol: 2.00+
  253. Boot loader hook (see ADVANCED BOOT LOADER HOOKS below.)
  254. Field name: start_sys_seg
  255. Type: read
  256. Offset/size: 0x20c/2
  257. Protocol: 2.00+
  258. The load low segment (0x1000). Obsolete.
  259. Field name: kernel_version
  260. Type: read
  261. Offset/size: 0x20e/2
  262. Protocol: 2.00+
  263. If set to a nonzero value, contains a pointer to a NUL-terminated
  264. human-readable kernel version number string, less 0x200. This can
  265. be used to display the kernel version to the user. This value
  266. should be less than (0x200*setup_sects).
  267. For example, if this value is set to 0x1c00, the kernel version
  268. number string can be found at offset 0x1e00 in the kernel file.
  269. This is a valid value if and only if the "setup_sects" field
  270. contains the value 15 or higher, as:
  271. 0x1c00 < 15*0x200 (= 0x1e00) but
  272. 0x1c00 >= 14*0x200 (= 0x1c00)
  273. 0x1c00 >> 9 = 14, so the minimum value for setup_secs is 15.
  274. Field name: type_of_loader
  275. Type: write (obligatory)
  276. Offset/size: 0x210/1
  277. Protocol: 2.00+
  278. If your boot loader has an assigned id (see table below), enter
  279. 0xTV here, where T is an identifier for the boot loader and V is
  280. a version number. Otherwise, enter 0xFF here.
  281. For boot loader IDs above T = 0xD, write T = 0xE to this field and
  282. write the extended ID minus 0x10 to the ext_loader_type field.
  283. Similarly, the ext_loader_ver field can be used to provide more than
  284. four bits for the bootloader version.
  285. For example, for T = 0x15, V = 0x234, write:
  286. type_of_loader <- 0xE4
  287. ext_loader_type <- 0x05
  288. ext_loader_ver <- 0x23
  289. Assigned boot loader ids:
  290. 0 LILO (0x00 reserved for pre-2.00 bootloader)
  291. 1 Loadlin
  292. 2 bootsect-loader (0x20, all other values reserved)
  293. 3 Syslinux
  294. 4 Etherboot/gPXE
  295. 5 ELILO
  296. 7 GRUB
  297. 8 U-Boot
  298. 9 Xen
  299. A Gujin
  300. B Qemu
  301. C Arcturus Networks uCbootloader
  302. E Extended (see ext_loader_type)
  303. F Special (0xFF = undefined)
  304. Please contact <hpa@zytor.com> if you need a bootloader ID
  305. value assigned.
  306. Field name: loadflags
  307. Type: modify (obligatory)
  308. Offset/size: 0x211/1
  309. Protocol: 2.00+
  310. This field is a bitmask.
  311. Bit 0 (read): LOADED_HIGH
  312. - If 0, the protected-mode code is loaded at 0x10000.
  313. - If 1, the protected-mode code is loaded at 0x100000.
  314. Bit 5 (write): QUIET_FLAG
  315. - If 0, print early messages.
  316. - If 1, suppress early messages.
  317. This requests to the kernel (decompressor and early
  318. kernel) to not write early messages that require
  319. accessing the display hardware directly.
  320. Bit 6 (write): KEEP_SEGMENTS
  321. Protocol: 2.07+
  322. - If 0, reload the segment registers in the 32bit entry point.
  323. - If 1, do not reload the segment registers in the 32bit entry point.
  324. Assume that %cs %ds %ss %es are all set to flat segments with
  325. a base of 0 (or the equivalent for their environment).
  326. Bit 7 (write): CAN_USE_HEAP
  327. Set this bit to 1 to indicate that the value entered in the
  328. heap_end_ptr is valid. If this field is clear, some setup code
  329. functionality will be disabled.
  330. Field name: setup_move_size
  331. Type: modify (obligatory)
  332. Offset/size: 0x212/2
  333. Protocol: 2.00-2.01
  334. When using protocol 2.00 or 2.01, if the real mode kernel is not
  335. loaded at 0x90000, it gets moved there later in the loading
  336. sequence. Fill in this field if you want additional data (such as
  337. the kernel command line) moved in addition to the real-mode kernel
  338. itself.
  339. The unit is bytes starting with the beginning of the boot sector.
  340. This field is can be ignored when the protocol is 2.02 or higher, or
  341. if the real-mode code is loaded at 0x90000.
  342. Field name: code32_start
  343. Type: modify (optional, reloc)
  344. Offset/size: 0x214/4
  345. Protocol: 2.00+
  346. The address to jump to in protected mode. This defaults to the load
  347. address of the kernel, and can be used by the boot loader to
  348. determine the proper load address.
  349. This field can be modified for two purposes:
  350. 1. as a boot loader hook (see ADVANCED BOOT LOADER HOOKS below.)
  351. 2. if a bootloader which does not install a hook loads a
  352. relocatable kernel at a nonstandard address it will have to modify
  353. this field to point to the load address.
  354. Field name: ramdisk_image
  355. Type: write (obligatory)
  356. Offset/size: 0x218/4
  357. Protocol: 2.00+
  358. The 32-bit linear address of the initial ramdisk or ramfs. Leave at
  359. zero if there is no initial ramdisk/ramfs.
  360. Field name: ramdisk_size
  361. Type: write (obligatory)
  362. Offset/size: 0x21c/4
  363. Protocol: 2.00+
  364. Size of the initial ramdisk or ramfs. Leave at zero if there is no
  365. initial ramdisk/ramfs.
  366. Field name: bootsect_kludge
  367. Type: kernel internal
  368. Offset/size: 0x220/4
  369. Protocol: 2.00+
  370. This field is obsolete.
  371. Field name: heap_end_ptr
  372. Type: write (obligatory)
  373. Offset/size: 0x224/2
  374. Protocol: 2.01+
  375. Set this field to the offset (from the beginning of the real-mode
  376. code) of the end of the setup stack/heap, minus 0x0200.
  377. Field name: ext_loader_ver
  378. Type: write (optional)
  379. Offset/size: 0x226/1
  380. Protocol: 2.02+
  381. This field is used as an extension of the version number in the
  382. type_of_loader field. The total version number is considered to be
  383. (type_of_loader & 0x0f) + (ext_loader_ver << 4).
  384. The use of this field is boot loader specific. If not written, it
  385. is zero.
  386. Kernels prior to 2.6.31 did not recognize this field, but it is safe
  387. to write for protocol version 2.02 or higher.
  388. Field name: ext_loader_type
  389. Type: write (obligatory if (type_of_loader & 0xf0) == 0xe0)
  390. Offset/size: 0x227/1
  391. Protocol: 2.02+
  392. This field is used as an extension of the type number in
  393. type_of_loader field. If the type in type_of_loader is 0xE, then
  394. the actual type is (ext_loader_type + 0x10).
  395. This field is ignored if the type in type_of_loader is not 0xE.
  396. Kernels prior to 2.6.31 did not recognize this field, but it is safe
  397. to write for protocol version 2.02 or higher.
  398. Field name: cmd_line_ptr
  399. Type: write (obligatory)
  400. Offset/size: 0x228/4
  401. Protocol: 2.02+
  402. Set this field to the linear address of the kernel command line.
  403. The kernel command line can be located anywhere between the end of
  404. the setup heap and 0xA0000; it does not have to be located in the
  405. same 64K segment as the real-mode code itself.
  406. Fill in this field even if your boot loader does not support a
  407. command line, in which case you can point this to an empty string
  408. (or better yet, to the string "auto".) If this field is left at
  409. zero, the kernel will assume that your boot loader does not support
  410. the 2.02+ protocol.
  411. Field name: ramdisk_max
  412. Type: read
  413. Offset/size: 0x22c/4
  414. Protocol: 2.03+
  415. The maximum address that may be occupied by the initial
  416. ramdisk/ramfs contents. For boot protocols 2.02 or earlier, this
  417. field is not present, and the maximum address is 0x37FFFFFF. (This
  418. address is defined as the address of the highest safe byte, so if
  419. your ramdisk is exactly 131072 bytes long and this field is
  420. 0x37FFFFFF, you can start your ramdisk at 0x37FE0000.)
  421. Field name: kernel_alignment
  422. Type: read/modify (reloc)
  423. Offset/size: 0x230/4
  424. Protocol: 2.05+ (read), 2.10+ (modify)
  425. Alignment unit required by the kernel (if relocatable_kernel is
  426. true.) A relocatable kernel that is loaded at an alignment
  427. incompatible with the value in this field will be realigned during
  428. kernel initialization.
  429. Starting with protocol version 2.10, this reflects the kernel
  430. alignment preferred for optimal performance; it is possible for the
  431. loader to modify this field to permit a lesser alignment. See the
  432. min_alignment and pref_address field below.
  433. Field name: relocatable_kernel
  434. Type: read (reloc)
  435. Offset/size: 0x234/1
  436. Protocol: 2.05+
  437. If this field is nonzero, the protected-mode part of the kernel can
  438. be loaded at any address that satisfies the kernel_alignment field.
  439. After loading, the boot loader must set the code32_start field to
  440. point to the loaded code, or to a boot loader hook.
  441. Field name: min_alignment
  442. Type: read (reloc)
  443. Offset/size: 0x235/1
  444. Protocol: 2.10+
  445. This field, if nonzero, indicates as a power of two the minimum
  446. alignment required, as opposed to preferred, by the kernel to boot.
  447. If a boot loader makes use of this field, it should update the
  448. kernel_alignment field with the alignment unit desired; typically:
  449. kernel_alignment = 1 << min_alignment
  450. There may be a considerable performance cost with an excessively
  451. misaligned kernel. Therefore, a loader should typically try each
  452. power-of-two alignment from kernel_alignment down to this alignment.
  453. Field name: cmdline_size
  454. Type: read
  455. Offset/size: 0x238/4
  456. Protocol: 2.06+
  457. The maximum size of the command line without the terminating
  458. zero. This means that the command line can contain at most
  459. cmdline_size characters. With protocol version 2.05 and earlier, the
  460. maximum size was 255.
  461. Field name: hardware_subarch
  462. Type: write (optional, defaults to x86/PC)
  463. Offset/size: 0x23c/4
  464. Protocol: 2.07+
  465. In a paravirtualized environment the hardware low level architectural
  466. pieces such as interrupt handling, page table handling, and
  467. accessing process control registers needs to be done differently.
  468. This field allows the bootloader to inform the kernel we are in one
  469. one of those environments.
  470. 0x00000000 The default x86/PC environment
  471. 0x00000001 lguest
  472. 0x00000002 Xen
  473. 0x00000003 Moorestown MID
  474. 0x00000004 CE4100 TV Platform
  475. Field name: hardware_subarch_data
  476. Type: write (subarch-dependent)
  477. Offset/size: 0x240/8
  478. Protocol: 2.07+
  479. A pointer to data that is specific to hardware subarch
  480. This field is currently unused for the default x86/PC environment,
  481. do not modify.
  482. Field name: payload_offset
  483. Type: read
  484. Offset/size: 0x248/4
  485. Protocol: 2.08+
  486. If non-zero then this field contains the offset from the beginning
  487. of the protected-mode code to the payload.
  488. The payload may be compressed. The format of both the compressed and
  489. uncompressed data should be determined using the standard magic
  490. numbers. The currently supported compression formats are gzip
  491. (magic numbers 1F 8B or 1F 9E), bzip2 (magic number 42 5A), LZMA
  492. (magic number 5D 00), and XZ (magic number FD 37). The uncompressed
  493. payload is currently always ELF (magic number 7F 45 4C 46).
  494. Field name: payload_length
  495. Type: read
  496. Offset/size: 0x24c/4
  497. Protocol: 2.08+
  498. The length of the payload.
  499. Field name: setup_data
  500. Type: write (special)
  501. Offset/size: 0x250/8
  502. Protocol: 2.09+
  503. The 64-bit physical pointer to NULL terminated single linked list of
  504. struct setup_data. This is used to define a more extensible boot
  505. parameters passing mechanism. The definition of struct setup_data is
  506. as follow:
  507. struct setup_data {
  508. u64 next;
  509. u32 type;
  510. u32 len;
  511. u8 data[0];
  512. };
  513. Where, the next is a 64-bit physical pointer to the next node of
  514. linked list, the next field of the last node is 0; the type is used
  515. to identify the contents of data; the len is the length of data
  516. field; the data holds the real payload.
  517. This list may be modified at a number of points during the bootup
  518. process. Therefore, when modifying this list one should always make
  519. sure to consider the case where the linked list already contains
  520. entries.
  521. Field name: pref_address
  522. Type: read (reloc)
  523. Offset/size: 0x258/8
  524. Protocol: 2.10+
  525. This field, if nonzero, represents a preferred load address for the
  526. kernel. A relocating bootloader should attempt to load at this
  527. address if possible.
  528. A non-relocatable kernel will unconditionally move itself and to run
  529. at this address.
  530. Field name: init_size
  531. Type: read
  532. Offset/size: 0x260/4
  533. This field indicates the amount of linear contiguous memory starting
  534. at the kernel runtime start address that the kernel needs before it
  535. is capable of examining its memory map. This is not the same thing
  536. as the total amount of memory the kernel needs to boot, but it can
  537. be used by a relocating boot loader to help select a safe load
  538. address for the kernel.
  539. The kernel runtime start address is determined by the following algorithm:
  540. if (relocatable_kernel)
  541. runtime_start = align_up(load_address, kernel_alignment)
  542. else
  543. runtime_start = pref_address
  544. **** THE IMAGE CHECKSUM
  545. From boot protocol version 2.08 onwards the CRC-32 is calculated over
  546. the entire file using the characteristic polynomial 0x04C11DB7 and an
  547. initial remainder of 0xffffffff. The checksum is appended to the
  548. file; therefore the CRC of the file up to the limit specified in the
  549. syssize field of the header is always 0.
  550. **** THE KERNEL COMMAND LINE
  551. The kernel command line has become an important way for the boot
  552. loader to communicate with the kernel. Some of its options are also
  553. relevant to the boot loader itself, see "special command line options"
  554. below.
  555. The kernel command line is a null-terminated string. The maximum
  556. length can be retrieved from the field cmdline_size. Before protocol
  557. version 2.06, the maximum was 255 characters. A string that is too
  558. long will be automatically truncated by the kernel.
  559. If the boot protocol version is 2.02 or later, the address of the
  560. kernel command line is given by the header field cmd_line_ptr (see
  561. above.) This address can be anywhere between the end of the setup
  562. heap and 0xA0000.
  563. If the protocol version is *not* 2.02 or higher, the kernel
  564. command line is entered using the following protocol:
  565. At offset 0x0020 (word), "cmd_line_magic", enter the magic
  566. number 0xA33F.
  567. At offset 0x0022 (word), "cmd_line_offset", enter the offset
  568. of the kernel command line (relative to the start of the
  569. real-mode kernel).
  570. The kernel command line *must* be within the memory region
  571. covered by setup_move_size, so you may need to adjust this
  572. field.
  573. **** MEMORY LAYOUT OF THE REAL-MODE CODE
  574. The real-mode code requires a stack/heap to be set up, as well as
  575. memory allocated for the kernel command line. This needs to be done
  576. in the real-mode accessible memory in bottom megabyte.
  577. It should be noted that modern machines often have a sizable Extended
  578. BIOS Data Area (EBDA). As a result, it is advisable to use as little
  579. of the low megabyte as possible.
  580. Unfortunately, under the following circumstances the 0x90000 memory
  581. segment has to be used:
  582. - When loading a zImage kernel ((loadflags & 0x01) == 0).
  583. - When loading a 2.01 or earlier boot protocol kernel.
  584. -> For the 2.00 and 2.01 boot protocols, the real-mode code
  585. can be loaded at another address, but it is internally
  586. relocated to 0x90000. For the "old" protocol, the
  587. real-mode code must be loaded at 0x90000.
  588. When loading at 0x90000, avoid using memory above 0x9a000.
  589. For boot protocol 2.02 or higher, the command line does not have to be
  590. located in the same 64K segment as the real-mode setup code; it is
  591. thus permitted to give the stack/heap the full 64K segment and locate
  592. the command line above it.
  593. The kernel command line should not be located below the real-mode
  594. code, nor should it be located in high memory.
  595. **** SAMPLE BOOT CONFIGURATION
  596. As a sample configuration, assume the following layout of the real
  597. mode segment:
  598. When loading below 0x90000, use the entire segment:
  599. 0x0000-0x7fff Real mode kernel
  600. 0x8000-0xdfff Stack and heap
  601. 0xe000-0xffff Kernel command line
  602. When loading at 0x90000 OR the protocol version is 2.01 or earlier:
  603. 0x0000-0x7fff Real mode kernel
  604. 0x8000-0x97ff Stack and heap
  605. 0x9800-0x9fff Kernel command line
  606. Such a boot loader should enter the following fields in the header:
  607. unsigned long base_ptr; /* base address for real-mode segment */
  608. if ( setup_sects == 0 ) {
  609. setup_sects = 4;
  610. }
  611. if ( protocol >= 0x0200 ) {
  612. type_of_loader = <type code>;
  613. if ( loading_initrd ) {
  614. ramdisk_image = <initrd_address>;
  615. ramdisk_size = <initrd_size>;
  616. }
  617. if ( protocol >= 0x0202 && loadflags & 0x01 )
  618. heap_end = 0xe000;
  619. else
  620. heap_end = 0x9800;
  621. if ( protocol >= 0x0201 ) {
  622. heap_end_ptr = heap_end - 0x200;
  623. loadflags |= 0x80; /* CAN_USE_HEAP */
  624. }
  625. if ( protocol >= 0x0202 ) {
  626. cmd_line_ptr = base_ptr + heap_end;
  627. strcpy(cmd_line_ptr, cmdline);
  628. } else {
  629. cmd_line_magic = 0xA33F;
  630. cmd_line_offset = heap_end;
  631. setup_move_size = heap_end + strlen(cmdline)+1;
  632. strcpy(base_ptr+cmd_line_offset, cmdline);
  633. }
  634. } else {
  635. /* Very old kernel */
  636. heap_end = 0x9800;
  637. cmd_line_magic = 0xA33F;
  638. cmd_line_offset = heap_end;
  639. /* A very old kernel MUST have its real-mode code
  640. loaded at 0x90000 */
  641. if ( base_ptr != 0x90000 ) {
  642. /* Copy the real-mode kernel */
  643. memcpy(0x90000, base_ptr, (setup_sects+1)*512);
  644. base_ptr = 0x90000; /* Relocated */
  645. }
  646. strcpy(0x90000+cmd_line_offset, cmdline);
  647. /* It is recommended to clear memory up to the 32K mark */
  648. memset(0x90000 + (setup_sects+1)*512, 0,
  649. (64-(setup_sects+1))*512);
  650. }
  651. **** LOADING THE REST OF THE KERNEL
  652. The 32-bit (non-real-mode) kernel starts at offset (setup_sects+1)*512
  653. in the kernel file (again, if setup_sects == 0 the real value is 4.)
  654. It should be loaded at address 0x10000 for Image/zImage kernels and
  655. 0x100000 for bzImage kernels.
  656. The kernel is a bzImage kernel if the protocol >= 2.00 and the 0x01
  657. bit (LOAD_HIGH) in the loadflags field is set:
  658. is_bzImage = (protocol >= 0x0200) && (loadflags & 0x01);
  659. load_address = is_bzImage ? 0x100000 : 0x10000;
  660. Note that Image/zImage kernels can be up to 512K in size, and thus use
  661. the entire 0x10000-0x90000 range of memory. This means it is pretty
  662. much a requirement for these kernels to load the real-mode part at
  663. 0x90000. bzImage kernels allow much more flexibility.
  664. **** SPECIAL COMMAND LINE OPTIONS
  665. If the command line provided by the boot loader is entered by the
  666. user, the user may expect the following command line options to work.
  667. They should normally not be deleted from the kernel command line even
  668. though not all of them are actually meaningful to the kernel. Boot
  669. loader authors who need additional command line options for the boot
  670. loader itself should get them registered in
  671. Documentation/kernel-parameters.txt to make sure they will not
  672. conflict with actual kernel options now or in the future.
  673. vga=<mode>
  674. <mode> here is either an integer (in C notation, either
  675. decimal, octal, or hexadecimal) or one of the strings
  676. "normal" (meaning 0xFFFF), "ext" (meaning 0xFFFE) or "ask"
  677. (meaning 0xFFFD). This value should be entered into the
  678. vid_mode field, as it is used by the kernel before the command
  679. line is parsed.
  680. mem=<size>
  681. <size> is an integer in C notation optionally followed by
  682. (case insensitive) K, M, G, T, P or E (meaning << 10, << 20,
  683. << 30, << 40, << 50 or << 60). This specifies the end of
  684. memory to the kernel. This affects the possible placement of
  685. an initrd, since an initrd should be placed near end of
  686. memory. Note that this is an option to *both* the kernel and
  687. the bootloader!
  688. initrd=<file>
  689. An initrd should be loaded. The meaning of <file> is
  690. obviously bootloader-dependent, and some boot loaders
  691. (e.g. LILO) do not have such a command.
  692. In addition, some boot loaders add the following options to the
  693. user-specified command line:
  694. BOOT_IMAGE=<file>
  695. The boot image which was loaded. Again, the meaning of <file>
  696. is obviously bootloader-dependent.
  697. auto
  698. The kernel was booted without explicit user intervention.
  699. If these options are added by the boot loader, it is highly
  700. recommended that they are located *first*, before the user-specified
  701. or configuration-specified command line. Otherwise, "init=/bin/sh"
  702. gets confused by the "auto" option.
  703. **** RUNNING THE KERNEL
  704. The kernel is started by jumping to the kernel entry point, which is
  705. located at *segment* offset 0x20 from the start of the real mode
  706. kernel. This means that if you loaded your real-mode kernel code at
  707. 0x90000, the kernel entry point is 9020:0000.
  708. At entry, ds = es = ss should point to the start of the real-mode
  709. kernel code (0x9000 if the code is loaded at 0x90000), sp should be
  710. set up properly, normally pointing to the top of the heap, and
  711. interrupts should be disabled. Furthermore, to guard against bugs in
  712. the kernel, it is recommended that the boot loader sets fs = gs = ds =
  713. es = ss.
  714. In our example from above, we would do:
  715. /* Note: in the case of the "old" kernel protocol, base_ptr must
  716. be == 0x90000 at this point; see the previous sample code */
  717. seg = base_ptr >> 4;
  718. cli(); /* Enter with interrupts disabled! */
  719. /* Set up the real-mode kernel stack */
  720. _SS = seg;
  721. _SP = heap_end;
  722. _DS = _ES = _FS = _GS = seg;
  723. jmp_far(seg+0x20, 0); /* Run the kernel */
  724. If your boot sector accesses a floppy drive, it is recommended to
  725. switch off the floppy motor before running the kernel, since the
  726. kernel boot leaves interrupts off and thus the motor will not be
  727. switched off, especially if the loaded kernel has the floppy driver as
  728. a demand-loaded module!
  729. **** ADVANCED BOOT LOADER HOOKS
  730. If the boot loader runs in a particularly hostile environment (such as
  731. LOADLIN, which runs under DOS) it may be impossible to follow the
  732. standard memory location requirements. Such a boot loader may use the
  733. following hooks that, if set, are invoked by the kernel at the
  734. appropriate time. The use of these hooks should probably be
  735. considered an absolutely last resort!
  736. IMPORTANT: All the hooks are required to preserve %esp, %ebp, %esi and
  737. %edi across invocation.
  738. realmode_swtch:
  739. A 16-bit real mode far subroutine invoked immediately before
  740. entering protected mode. The default routine disables NMI, so
  741. your routine should probably do so, too.
  742. code32_start:
  743. A 32-bit flat-mode routine *jumped* to immediately after the
  744. transition to protected mode, but before the kernel is
  745. uncompressed. No segments, except CS, are guaranteed to be
  746. set up (current kernels do, but older ones do not); you should
  747. set them up to BOOT_DS (0x18) yourself.
  748. After completing your hook, you should jump to the address
  749. that was in this field before your boot loader overwrote it
  750. (relocated, if appropriate.)
  751. **** 32-bit BOOT PROTOCOL
  752. For machine with some new BIOS other than legacy BIOS, such as EFI,
  753. LinuxBIOS, etc, and kexec, the 16-bit real mode setup code in kernel
  754. based on legacy BIOS can not be used, so a 32-bit boot protocol needs
  755. to be defined.
  756. In 32-bit boot protocol, the first step in loading a Linux kernel
  757. should be to setup the boot parameters (struct boot_params,
  758. traditionally known as "zero page"). The memory for struct boot_params
  759. should be allocated and initialized to all zero. Then the setup header
  760. from offset 0x01f1 of kernel image on should be loaded into struct
  761. boot_params and examined. The end of setup header can be calculated as
  762. follow:
  763. 0x0202 + byte value at offset 0x0201
  764. In addition to read/modify/write the setup header of the struct
  765. boot_params as that of 16-bit boot protocol, the boot loader should
  766. also fill the additional fields of the struct boot_params as that
  767. described in zero-page.txt.
  768. After setupping the struct boot_params, the boot loader can load the
  769. 32/64-bit kernel in the same way as that of 16-bit boot protocol.
  770. In 32-bit boot protocol, the kernel is started by jumping to the
  771. 32-bit kernel entry point, which is the start address of loaded
  772. 32/64-bit kernel.
  773. At entry, the CPU must be in 32-bit protected mode with paging
  774. disabled; a GDT must be loaded with the descriptors for selectors
  775. __BOOT_CS(0x10) and __BOOT_DS(0x18); both descriptors must be 4G flat
  776. segment; __BOOS_CS must have execute/read permission, and __BOOT_DS
  777. must have read/write permission; CS must be __BOOT_CS and DS, ES, SS
  778. must be __BOOT_DS; interrupt must be disabled; %esi must hold the base
  779. address of the struct boot_params; %ebp, %edi and %ebx must be zero.