boot.txt 29 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853
  1. THE LINUX/I386 BOOT PROTOCOL
  2. ----------------------------
  3. H. Peter Anvin <hpa@zytor.com>
  4. Last update 2007-05-23
  5. On the i386 platform, the Linux kernel uses a rather complicated boot
  6. convention. This has evolved partially due to historical aspects, as
  7. well as the desire in the early days to have the kernel itself be a
  8. bootable image, the complicated PC memory model and due to changed
  9. expectations in the PC industry caused by the effective demise of
  10. real-mode DOS as a mainstream operating system.
  11. Currently, the following versions of the Linux/i386 boot protocol exist.
  12. Old kernels: zImage/Image support only. Some very early kernels
  13. may not even support a command line.
  14. Protocol 2.00: (Kernel 1.3.73) Added bzImage and initrd support, as
  15. well as a formalized way to communicate between the
  16. boot loader and the kernel. setup.S made relocatable,
  17. although the traditional setup area still assumed
  18. writable.
  19. Protocol 2.01: (Kernel 1.3.76) Added a heap overrun warning.
  20. Protocol 2.02: (Kernel 2.4.0-test3-pre3) New command line protocol.
  21. Lower the conventional memory ceiling. No overwrite
  22. of the traditional setup area, thus making booting
  23. safe for systems which use the EBDA from SMM or 32-bit
  24. BIOS entry points. zImage deprecated but still
  25. supported.
  26. Protocol 2.03: (Kernel 2.4.18-pre1) Explicitly makes the highest possible
  27. initrd address available to the bootloader.
  28. Protocol 2.04: (Kernel 2.6.14) Extend the syssize field to four bytes.
  29. Protocol 2.05: (Kernel 2.6.20) Make protected mode kernel relocatable.
  30. Introduce relocatable_kernel and kernel_alignment fields.
  31. Protocol 2.06: (Kernel 2.6.22) Added a field that contains the size of
  32. the boot command line
  33. **** MEMORY LAYOUT
  34. The traditional memory map for the kernel loader, used for Image or
  35. zImage kernels, typically looks like:
  36. | |
  37. 0A0000 +------------------------+
  38. | Reserved for BIOS | Do not use. Reserved for BIOS EBDA.
  39. 09A000 +------------------------+
  40. | Command line |
  41. | Stack/heap | For use by the kernel real-mode code.
  42. 098000 +------------------------+
  43. | Kernel setup | The kernel real-mode code.
  44. 090200 +------------------------+
  45. | Kernel boot sector | The kernel legacy boot sector.
  46. 090000 +------------------------+
  47. | Protected-mode kernel | The bulk of the kernel image.
  48. 010000 +------------------------+
  49. | Boot loader | <- Boot sector entry point 0000:7C00
  50. 001000 +------------------------+
  51. | Reserved for MBR/BIOS |
  52. 000800 +------------------------+
  53. | Typically used by MBR |
  54. 000600 +------------------------+
  55. | BIOS use only |
  56. 000000 +------------------------+
  57. When using bzImage, the protected-mode kernel was relocated to
  58. 0x100000 ("high memory"), and the kernel real-mode block (boot sector,
  59. setup, and stack/heap) was made relocatable to any address between
  60. 0x10000 and end of low memory. Unfortunately, in protocols 2.00 and
  61. 2.01 the 0x90000+ memory range is still used internally by the kernel;
  62. the 2.02 protocol resolves that problem.
  63. It is desirable to keep the "memory ceiling" -- the highest point in
  64. low memory touched by the boot loader -- as low as possible, since
  65. some newer BIOSes have begun to allocate some rather large amounts of
  66. memory, called the Extended BIOS Data Area, near the top of low
  67. memory. The boot loader should use the "INT 12h" BIOS call to verify
  68. how much low memory is available.
  69. Unfortunately, if INT 12h reports that the amount of memory is too
  70. low, there is usually nothing the boot loader can do but to report an
  71. error to the user. The boot loader should therefore be designed to
  72. take up as little space in low memory as it reasonably can. For
  73. zImage or old bzImage kernels, which need data written into the
  74. 0x90000 segment, the boot loader should make sure not to use memory
  75. above the 0x9A000 point; too many BIOSes will break above that point.
  76. For a modern bzImage kernel with boot protocol version >= 2.02, a
  77. memory layout like the following is suggested:
  78. ~ ~
  79. | Protected-mode kernel |
  80. 100000 +------------------------+
  81. | I/O memory hole |
  82. 0A0000 +------------------------+
  83. | Reserved for BIOS | Leave as much as possible unused
  84. ~ ~
  85. | Command line | (Can also be below the X+10000 mark)
  86. X+10000 +------------------------+
  87. | Stack/heap | For use by the kernel real-mode code.
  88. X+08000 +------------------------+
  89. | Kernel setup | The kernel real-mode code.
  90. | Kernel boot sector | The kernel legacy boot sector.
  91. X +------------------------+
  92. | Boot loader | <- Boot sector entry point 0000:7C00
  93. 001000 +------------------------+
  94. | Reserved for MBR/BIOS |
  95. 000800 +------------------------+
  96. | Typically used by MBR |
  97. 000600 +------------------------+
  98. | BIOS use only |
  99. 000000 +------------------------+
  100. ... where the address X is as low as the design of the boot loader
  101. permits.
  102. **** THE REAL-MODE KERNEL HEADER
  103. In the following text, and anywhere in the kernel boot sequence, "a
  104. sector" refers to 512 bytes. It is independent of the actual sector
  105. size of the underlying medium.
  106. The first step in loading a Linux kernel should be to load the
  107. real-mode code (boot sector and setup code) and then examine the
  108. following header at offset 0x01f1. The real-mode code can total up to
  109. 32K, although the boot loader may choose to load only the first two
  110. sectors (1K) and then examine the bootup sector size.
  111. The header looks like:
  112. Offset Proto Name Meaning
  113. /Size
  114. 01F1/1 ALL(1 setup_sects The size of the setup in sectors
  115. 01F2/2 ALL root_flags If set, the root is mounted readonly
  116. 01F4/4 2.04+(2 syssize The size of the 32-bit code in 16-byte paras
  117. 01F8/2 ALL ram_size DO NOT USE - for bootsect.S use only
  118. 01FA/2 ALL vid_mode Video mode control
  119. 01FC/2 ALL root_dev Default root device number
  120. 01FE/2 ALL boot_flag 0xAA55 magic number
  121. 0200/2 2.00+ jump Jump instruction
  122. 0202/4 2.00+ header Magic signature "HdrS"
  123. 0206/2 2.00+ version Boot protocol version supported
  124. 0208/4 2.00+ realmode_swtch Boot loader hook (see below)
  125. 020C/2 2.00+ start_sys The load-low segment (0x1000) (obsolete)
  126. 020E/2 2.00+ kernel_version Pointer to kernel version string
  127. 0210/1 2.00+ type_of_loader Boot loader identifier
  128. 0211/1 2.00+ loadflags Boot protocol option flags
  129. 0212/2 2.00+ setup_move_size Move to high memory size (used with hooks)
  130. 0214/4 2.00+ code32_start Boot loader hook (see below)
  131. 0218/4 2.00+ ramdisk_image initrd load address (set by boot loader)
  132. 021C/4 2.00+ ramdisk_size initrd size (set by boot loader)
  133. 0220/4 2.00+ bootsect_kludge DO NOT USE - for bootsect.S use only
  134. 0224/2 2.01+ heap_end_ptr Free memory after setup end
  135. 0226/2 N/A pad1 Unused
  136. 0228/4 2.02+ cmd_line_ptr 32-bit pointer to the kernel command line
  137. 022C/4 2.03+ initrd_addr_max Highest legal initrd address
  138. 0230/4 2.05+ kernel_alignment Physical addr alignment required for kernel
  139. 0234/1 2.05+ relocatable_kernel Whether kernel is relocatable or not
  140. 0235/3 N/A pad2 Unused
  141. 0238/4 2.06+ cmdline_size Maximum size of the kernel command line
  142. 023C/4 2.07+ hardware_subarch Hardware subarchitecture
  143. 0240/8 2.07+ hardware_subarch_data Subarchitecture-specific data
  144. 0248/4 2.08+ payload_offset Offset of kernel payload
  145. 024C/4 2.08+ payload_length Length of kernel payload
  146. (1) For backwards compatibility, if the setup_sects field contains 0, the
  147. real value is 4.
  148. (2) For boot protocol prior to 2.04, the upper two bytes of the syssize
  149. field are unusable, which means the size of a bzImage kernel
  150. cannot be determined.
  151. If the "HdrS" (0x53726448) magic number is not found at offset 0x202,
  152. the boot protocol version is "old". Loading an old kernel, the
  153. following parameters should be assumed:
  154. Image type = zImage
  155. initrd not supported
  156. Real-mode kernel must be located at 0x90000.
  157. Otherwise, the "version" field contains the protocol version,
  158. e.g. protocol version 2.01 will contain 0x0201 in this field. When
  159. setting fields in the header, you must make sure only to set fields
  160. supported by the protocol version in use.
  161. **** DETAILS OF HEADER FIELDS
  162. For each field, some are information from the kernel to the bootloader
  163. ("read"), some are expected to be filled out by the bootloader
  164. ("write"), and some are expected to be read and modified by the
  165. bootloader ("modify").
  166. All general purpose boot loaders should write the fields marked
  167. (obligatory). Boot loaders who want to load the kernel at a
  168. nonstandard address should fill in the fields marked (reloc); other
  169. boot loaders can ignore those fields.
  170. The byte order of all fields is littleendian (this is x86, after all.)
  171. Field name: setup_sects
  172. Type: read
  173. Offset/size: 0x1f1/1
  174. Protocol: ALL
  175. The size of the setup code in 512-byte sectors. If this field is
  176. 0, the real value is 4. The real-mode code consists of the boot
  177. sector (always one 512-byte sector) plus the setup code.
  178. Field name: root_flags
  179. Type: modify (optional)
  180. Offset/size: 0x1f2/2
  181. Protocol: ALL
  182. If this field is nonzero, the root defaults to readonly. The use of
  183. this field is deprecated; use the "ro" or "rw" options on the
  184. command line instead.
  185. Field name: syssize
  186. Type: read
  187. Offset/size: 0x1f4/4 (protocol 2.04+) 0x1f4/2 (protocol ALL)
  188. Protocol: 2.04+
  189. The size of the protected-mode code in units of 16-byte paragraphs.
  190. For protocol versions older than 2.04 this field is only two bytes
  191. wide, and therefore cannot be trusted for the size of a kernel if
  192. the LOAD_HIGH flag is set.
  193. Field name: ram_size
  194. Type: kernel internal
  195. Offset/size: 0x1f8/2
  196. Protocol: ALL
  197. This field is obsolete.
  198. Field name: vid_mode
  199. Type: modify (obligatory)
  200. Offset/size: 0x1fa/2
  201. Please see the section on SPECIAL COMMAND LINE OPTIONS.
  202. Field name: root_dev
  203. Type: modify (optional)
  204. Offset/size: 0x1fc/2
  205. Protocol: ALL
  206. The default root device device number. The use of this field is
  207. deprecated, use the "root=" option on the command line instead.
  208. Field name: boot_flag
  209. Type: read
  210. Offset/size: 0x1fe/2
  211. Protocol: ALL
  212. Contains 0xAA55. This is the closest thing old Linux kernels have
  213. to a magic number.
  214. Field name: jump
  215. Type: read
  216. Offset/size: 0x200/2
  217. Protocol: 2.00+
  218. Contains an x86 jump instruction, 0xEB followed by a signed offset
  219. relative to byte 0x202. This can be used to determine the size of
  220. the header.
  221. Field name: header
  222. Type: read
  223. Offset/size: 0x202/4
  224. Protocol: 2.00+
  225. Contains the magic number "HdrS" (0x53726448).
  226. Field name: version
  227. Type: read
  228. Offset/size: 0x206/2
  229. Protocol: 2.00+
  230. Contains the boot protocol version, in (major << 8)+minor format,
  231. e.g. 0x0204 for version 2.04, and 0x0a11 for a hypothetical version
  232. 10.17.
  233. Field name: readmode_swtch
  234. Type: modify (optional)
  235. Offset/size: 0x208/4
  236. Protocol: 2.00+
  237. Boot loader hook (see ADVANCED BOOT LOADER HOOKS below.)
  238. Field name: start_sys
  239. Type: read
  240. Offset/size: 0x20c/4
  241. Protocol: 2.00+
  242. The load low segment (0x1000). Obsolete.
  243. Field name: kernel_version
  244. Type: read
  245. Offset/size: 0x20e/2
  246. Protocol: 2.00+
  247. If set to a nonzero value, contains a pointer to a NUL-terminated
  248. human-readable kernel version number string, less 0x200. This can
  249. be used to display the kernel version to the user. This value
  250. should be less than (0x200*setup_sects).
  251. For example, if this value is set to 0x1c00, the kernel version
  252. number string can be found at offset 0x1e00 in the kernel file.
  253. This is a valid value if and only if the "setup_sects" field
  254. contains the value 15 or higher, as:
  255. 0x1c00 < 15*0x200 (= 0x1e00) but
  256. 0x1c00 >= 14*0x200 (= 0x1c00)
  257. 0x1c00 >> 9 = 14, so the minimum value for setup_secs is 15.
  258. Field name: type_of_loader
  259. Type: write (obligatory)
  260. Offset/size: 0x210/1
  261. Protocol: 2.00+
  262. If your boot loader has an assigned id (see table below), enter
  263. 0xTV here, where T is an identifier for the boot loader and V is
  264. a version number. Otherwise, enter 0xFF here.
  265. Assigned boot loader ids:
  266. 0 LILO (0x00 reserved for pre-2.00 bootloader)
  267. 1 Loadlin
  268. 2 bootsect-loader (0x20, all other values reserved)
  269. 3 SYSLINUX
  270. 4 EtherBoot
  271. 5 ELILO
  272. 7 GRuB
  273. 8 U-BOOT
  274. 9 Xen
  275. A Gujin
  276. B Qemu
  277. Please contact <hpa@zytor.com> if you need a bootloader ID
  278. value assigned.
  279. Field name: loadflags
  280. Type: modify (obligatory)
  281. Offset/size: 0x211/1
  282. Protocol: 2.00+
  283. This field is a bitmask.
  284. Bit 0 (read): LOADED_HIGH
  285. - If 0, the protected-mode code is loaded at 0x10000.
  286. - If 1, the protected-mode code is loaded at 0x100000.
  287. Bit 6 (write): KEEP_SEGMENTS
  288. Protocol: 2.07+
  289. - if 0, reload the segment registers in the 32bit entry point.
  290. - if 1, do not reload the segment registers in the 32bit entry point.
  291. Assume that %cs %ds %ss %es are all set to flat segments with
  292. a base of 0 (or the equivalent for their environment).
  293. Bit 7 (write): CAN_USE_HEAP
  294. Set this bit to 1 to indicate that the value entered in the
  295. heap_end_ptr is valid. If this field is clear, some setup code
  296. functionality will be disabled.
  297. Field name: setup_move_size
  298. Type: modify (obligatory)
  299. Offset/size: 0x212/2
  300. Protocol: 2.00-2.01
  301. When using protocol 2.00 or 2.01, if the real mode kernel is not
  302. loaded at 0x90000, it gets moved there later in the loading
  303. sequence. Fill in this field if you want additional data (such as
  304. the kernel command line) moved in addition to the real-mode kernel
  305. itself.
  306. The unit is bytes starting with the beginning of the boot sector.
  307. This field is can be ignored when the protocol is 2.02 or higher, or
  308. if the real-mode code is loaded at 0x90000.
  309. Field name: code32_start
  310. Type: modify (optional, reloc)
  311. Offset/size: 0x214/4
  312. Protocol: 2.00+
  313. The address to jump to in protected mode. This defaults to the load
  314. address of the kernel, and can be used by the boot loader to
  315. determine the proper load address.
  316. This field can be modified for two purposes:
  317. 1. as a boot loader hook (see ADVANCED BOOT LOADER HOOKS below.)
  318. 2. if a bootloader which does not install a hook loads a
  319. relocatable kernel at a nonstandard address it will have to modify
  320. this field to point to the load address.
  321. Field name: ramdisk_image
  322. Type: write (obligatory)
  323. Offset/size: 0x218/4
  324. Protocol: 2.00+
  325. The 32-bit linear address of the initial ramdisk or ramfs. Leave at
  326. zero if there is no initial ramdisk/ramfs.
  327. Field name: ramdisk_size
  328. Type: write (obligatory)
  329. Offset/size: 0x21c/4
  330. Protocol: 2.00+
  331. Size of the initial ramdisk or ramfs. Leave at zero if there is no
  332. initial ramdisk/ramfs.
  333. Field name: bootsect_kludge
  334. Type: kernel internal
  335. Offset/size: 0x220/4
  336. Protocol: 2.00+
  337. This field is obsolete.
  338. Field name: heap_end_ptr
  339. Type: write (obligatory)
  340. Offset/size: 0x224/2
  341. Protocol: 2.01+
  342. Set this field to the offset (from the beginning of the real-mode
  343. code) of the end of the setup stack/heap, minus 0x0200.
  344. Field name: cmd_line_ptr
  345. Type: write (obligatory)
  346. Offset/size: 0x228/4
  347. Protocol: 2.02+
  348. Set this field to the linear address of the kernel command line.
  349. The kernel command line can be located anywhere between the end of
  350. the setup heap and 0xA0000; it does not have to be located in the
  351. same 64K segment as the real-mode code itself.
  352. Fill in this field even if your boot loader does not support a
  353. command line, in which case you can point this to an empty string
  354. (or better yet, to the string "auto".) If this field is left at
  355. zero, the kernel will assume that your boot loader does not support
  356. the 2.02+ protocol.
  357. Field name: initrd_addr_max
  358. Type: read
  359. Offset/size: 0x22c/4
  360. Protocol: 2.03+
  361. The maximum address that may be occupied by the initial
  362. ramdisk/ramfs contents. For boot protocols 2.02 or earlier, this
  363. field is not present, and the maximum address is 0x37FFFFFF. (This
  364. address is defined as the address of the highest safe byte, so if
  365. your ramdisk is exactly 131072 bytes long and this field is
  366. 0x37FFFFFF, you can start your ramdisk at 0x37FE0000.)
  367. Field name: kernel_alignment
  368. Type: read (reloc)
  369. Offset/size: 0x230/4
  370. Protocol: 2.05+
  371. Alignment unit required by the kernel (if relocatable_kernel is true.)
  372. Field name: relocatable_kernel
  373. Type: read (reloc)
  374. Offset/size: 0x234/1
  375. Protocol: 2.05+
  376. If this field is nonzero, the protected-mode part of the kernel can
  377. be loaded at any address that satisfies the kernel_alignment field.
  378. After loading, the boot loader must set the code32_start field to
  379. point to the loaded code, or to a boot loader hook.
  380. Field name: cmdline_size
  381. Type: read
  382. Offset/size: 0x238/4
  383. Protocol: 2.06+
  384. The maximum size of the command line without the terminating
  385. zero. This means that the command line can contain at most
  386. cmdline_size characters. With protocol version 2.05 and earlier, the
  387. maximum size was 255.
  388. Field name: hardware_subarch
  389. Type: write
  390. Offset/size: 0x23c/4
  391. Protocol: 2.07+
  392. In a paravirtualized environment the hardware low level architectural
  393. pieces such as interrupt handling, page table handling, and
  394. accessing process control registers needs to be done differently.
  395. This field allows the bootloader to inform the kernel we are in one
  396. one of those environments.
  397. 0x00000000 The default x86/PC environment
  398. 0x00000001 lguest
  399. 0x00000002 Xen
  400. Field name: hardware_subarch_data
  401. Type: write
  402. Offset/size: 0x240/8
  403. Protocol: 2.07+
  404. A pointer to data that is specific to hardware subarch
  405. Field name: payload_offset
  406. Type: read
  407. Offset/size: 0x248/4
  408. Protocol: 2.08+
  409. If non-zero then this field contains the offset from the end of the
  410. real-mode code to the payload.
  411. The payload may be compressed. The format of both the compressed and
  412. uncompressed data should be determined using the standard magic
  413. numbers. Currently only gzip compressed ELF is used.
  414. Field name: payload_length
  415. Type: read
  416. Offset/size: 0x24c/4
  417. Protocol: 2.08+
  418. The length of the payload.
  419. **** THE IMAGE CHECKSUM
  420. From boot protocol version 2.08 onwards the CRC-32 is calculated over
  421. the entire file using the characteristic polynomial 0x04C11DB7 and an
  422. initial remainder of 0xffffffff. The checksum is appended to the
  423. file; therefore the CRC of the file up to the limit specified in the
  424. syssize field of the header is always 0.
  425. **** THE KERNEL COMMAND LINE
  426. The kernel command line has become an important way for the boot
  427. loader to communicate with the kernel. Some of its options are also
  428. relevant to the boot loader itself, see "special command line options"
  429. below.
  430. The kernel command line is a null-terminated string. The maximum
  431. length can be retrieved from the field cmdline_size. Before protocol
  432. version 2.06, the maximum was 255 characters. A string that is too
  433. long will be automatically truncated by the kernel.
  434. If the boot protocol version is 2.02 or later, the address of the
  435. kernel command line is given by the header field cmd_line_ptr (see
  436. above.) This address can be anywhere between the end of the setup
  437. heap and 0xA0000.
  438. If the protocol version is *not* 2.02 or higher, the kernel
  439. command line is entered using the following protocol:
  440. At offset 0x0020 (word), "cmd_line_magic", enter the magic
  441. number 0xA33F.
  442. At offset 0x0022 (word), "cmd_line_offset", enter the offset
  443. of the kernel command line (relative to the start of the
  444. real-mode kernel).
  445. The kernel command line *must* be within the memory region
  446. covered by setup_move_size, so you may need to adjust this
  447. field.
  448. **** MEMORY LAYOUT OF THE REAL-MODE CODE
  449. The real-mode code requires a stack/heap to be set up, as well as
  450. memory allocated for the kernel command line. This needs to be done
  451. in the real-mode accessible memory in bottom megabyte.
  452. It should be noted that modern machines often have a sizable Extended
  453. BIOS Data Area (EBDA). As a result, it is advisable to use as little
  454. of the low megabyte as possible.
  455. Unfortunately, under the following circumstances the 0x90000 memory
  456. segment has to be used:
  457. - When loading a zImage kernel ((loadflags & 0x01) == 0).
  458. - When loading a 2.01 or earlier boot protocol kernel.
  459. -> For the 2.00 and 2.01 boot protocols, the real-mode code
  460. can be loaded at another address, but it is internally
  461. relocated to 0x90000. For the "old" protocol, the
  462. real-mode code must be loaded at 0x90000.
  463. When loading at 0x90000, avoid using memory above 0x9a000.
  464. For boot protocol 2.02 or higher, the command line does not have to be
  465. located in the same 64K segment as the real-mode setup code; it is
  466. thus permitted to give the stack/heap the full 64K segment and locate
  467. the command line above it.
  468. The kernel command line should not be located below the real-mode
  469. code, nor should it be located in high memory.
  470. **** SAMPLE BOOT CONFIGURATION
  471. As a sample configuration, assume the following layout of the real
  472. mode segment:
  473. When loading below 0x90000, use the entire segment:
  474. 0x0000-0x7fff Real mode kernel
  475. 0x8000-0xdfff Stack and heap
  476. 0xe000-0xffff Kernel command line
  477. When loading at 0x90000 OR the protocol version is 2.01 or earlier:
  478. 0x0000-0x7fff Real mode kernel
  479. 0x8000-0x97ff Stack and heap
  480. 0x9800-0x9fff Kernel command line
  481. Such a boot loader should enter the following fields in the header:
  482. unsigned long base_ptr; /* base address for real-mode segment */
  483. if ( setup_sects == 0 ) {
  484. setup_sects = 4;
  485. }
  486. if ( protocol >= 0x0200 ) {
  487. type_of_loader = <type code>;
  488. if ( loading_initrd ) {
  489. ramdisk_image = <initrd_address>;
  490. ramdisk_size = <initrd_size>;
  491. }
  492. if ( protocol >= 0x0202 && loadflags & 0x01 )
  493. heap_end = 0xe000;
  494. else
  495. heap_end = 0x9800;
  496. if ( protocol >= 0x0201 ) {
  497. heap_end_ptr = heap_end - 0x200;
  498. loadflags |= 0x80; /* CAN_USE_HEAP */
  499. }
  500. if ( protocol >= 0x0202 ) {
  501. cmd_line_ptr = base_ptr + heap_end;
  502. strcpy(cmd_line_ptr, cmdline);
  503. } else {
  504. cmd_line_magic = 0xA33F;
  505. cmd_line_offset = heap_end;
  506. setup_move_size = heap_end + strlen(cmdline)+1;
  507. strcpy(base_ptr+cmd_line_offset, cmdline);
  508. }
  509. } else {
  510. /* Very old kernel */
  511. heap_end = 0x9800;
  512. cmd_line_magic = 0xA33F;
  513. cmd_line_offset = heap_end;
  514. /* A very old kernel MUST have its real-mode code
  515. loaded at 0x90000 */
  516. if ( base_ptr != 0x90000 ) {
  517. /* Copy the real-mode kernel */
  518. memcpy(0x90000, base_ptr, (setup_sects+1)*512);
  519. base_ptr = 0x90000; /* Relocated */
  520. }
  521. strcpy(0x90000+cmd_line_offset, cmdline);
  522. /* It is recommended to clear memory up to the 32K mark */
  523. memset(0x90000 + (setup_sects+1)*512, 0,
  524. (64-(setup_sects+1))*512);
  525. }
  526. **** LOADING THE REST OF THE KERNEL
  527. The 32-bit (non-real-mode) kernel starts at offset (setup_sects+1)*512
  528. in the kernel file (again, if setup_sects == 0 the real value is 4.)
  529. It should be loaded at address 0x10000 for Image/zImage kernels and
  530. 0x100000 for bzImage kernels.
  531. The kernel is a bzImage kernel if the protocol >= 2.00 and the 0x01
  532. bit (LOAD_HIGH) in the loadflags field is set:
  533. is_bzImage = (protocol >= 0x0200) && (loadflags & 0x01);
  534. load_address = is_bzImage ? 0x100000 : 0x10000;
  535. Note that Image/zImage kernels can be up to 512K in size, and thus use
  536. the entire 0x10000-0x90000 range of memory. This means it is pretty
  537. much a requirement for these kernels to load the real-mode part at
  538. 0x90000. bzImage kernels allow much more flexibility.
  539. **** SPECIAL COMMAND LINE OPTIONS
  540. If the command line provided by the boot loader is entered by the
  541. user, the user may expect the following command line options to work.
  542. They should normally not be deleted from the kernel command line even
  543. though not all of them are actually meaningful to the kernel. Boot
  544. loader authors who need additional command line options for the boot
  545. loader itself should get them registered in
  546. Documentation/kernel-parameters.txt to make sure they will not
  547. conflict with actual kernel options now or in the future.
  548. vga=<mode>
  549. <mode> here is either an integer (in C notation, either
  550. decimal, octal, or hexadecimal) or one of the strings
  551. "normal" (meaning 0xFFFF), "ext" (meaning 0xFFFE) or "ask"
  552. (meaning 0xFFFD). This value should be entered into the
  553. vid_mode field, as it is used by the kernel before the command
  554. line is parsed.
  555. mem=<size>
  556. <size> is an integer in C notation optionally followed by
  557. (case insensitive) K, M, G, T, P or E (meaning << 10, << 20,
  558. << 30, << 40, << 50 or << 60). This specifies the end of
  559. memory to the kernel. This affects the possible placement of
  560. an initrd, since an initrd should be placed near end of
  561. memory. Note that this is an option to *both* the kernel and
  562. the bootloader!
  563. initrd=<file>
  564. An initrd should be loaded. The meaning of <file> is
  565. obviously bootloader-dependent, and some boot loaders
  566. (e.g. LILO) do not have such a command.
  567. In addition, some boot loaders add the following options to the
  568. user-specified command line:
  569. BOOT_IMAGE=<file>
  570. The boot image which was loaded. Again, the meaning of <file>
  571. is obviously bootloader-dependent.
  572. auto
  573. The kernel was booted without explicit user intervention.
  574. If these options are added by the boot loader, it is highly
  575. recommended that they are located *first*, before the user-specified
  576. or configuration-specified command line. Otherwise, "init=/bin/sh"
  577. gets confused by the "auto" option.
  578. **** RUNNING THE KERNEL
  579. The kernel is started by jumping to the kernel entry point, which is
  580. located at *segment* offset 0x20 from the start of the real mode
  581. kernel. This means that if you loaded your real-mode kernel code at
  582. 0x90000, the kernel entry point is 9020:0000.
  583. At entry, ds = es = ss should point to the start of the real-mode
  584. kernel code (0x9000 if the code is loaded at 0x90000), sp should be
  585. set up properly, normally pointing to the top of the heap, and
  586. interrupts should be disabled. Furthermore, to guard against bugs in
  587. the kernel, it is recommended that the boot loader sets fs = gs = ds =
  588. es = ss.
  589. In our example from above, we would do:
  590. /* Note: in the case of the "old" kernel protocol, base_ptr must
  591. be == 0x90000 at this point; see the previous sample code */
  592. seg = base_ptr >> 4;
  593. cli(); /* Enter with interrupts disabled! */
  594. /* Set up the real-mode kernel stack */
  595. _SS = seg;
  596. _SP = heap_end;
  597. _DS = _ES = _FS = _GS = seg;
  598. jmp_far(seg+0x20, 0); /* Run the kernel */
  599. If your boot sector accesses a floppy drive, it is recommended to
  600. switch off the floppy motor before running the kernel, since the
  601. kernel boot leaves interrupts off and thus the motor will not be
  602. switched off, especially if the loaded kernel has the floppy driver as
  603. a demand-loaded module!
  604. **** ADVANCED BOOT LOADER HOOKS
  605. If the boot loader runs in a particularly hostile environment (such as
  606. LOADLIN, which runs under DOS) it may be impossible to follow the
  607. standard memory location requirements. Such a boot loader may use the
  608. following hooks that, if set, are invoked by the kernel at the
  609. appropriate time. The use of these hooks should probably be
  610. considered an absolutely last resort!
  611. IMPORTANT: All the hooks are required to preserve %esp, %ebp, %esi and
  612. %edi across invocation.
  613. realmode_swtch:
  614. A 16-bit real mode far subroutine invoked immediately before
  615. entering protected mode. The default routine disables NMI, so
  616. your routine should probably do so, too.
  617. code32_start:
  618. A 32-bit flat-mode routine *jumped* to immediately after the
  619. transition to protected mode, but before the kernel is
  620. uncompressed. No segments, except CS, are guaranteed to be
  621. set up (current kernels do, but older ones do not); you should
  622. set them up to BOOT_DS (0x18) yourself.
  623. After completing your hook, you should jump to the address
  624. that was in this field before your boot loader overwrote it
  625. (relocated, if appropriate.)
  626. **** 32-bit BOOT PROTOCOL
  627. For machine with some new BIOS other than legacy BIOS, such as EFI,
  628. LinuxBIOS, etc, and kexec, the 16-bit real mode setup code in kernel
  629. based on legacy BIOS can not be used, so a 32-bit boot protocol needs
  630. to be defined.
  631. In 32-bit boot protocol, the first step in loading a Linux kernel
  632. should be to setup the boot parameters (struct boot_params,
  633. traditionally known as "zero page"). The memory for struct boot_params
  634. should be allocated and initialized to all zero. Then the setup header
  635. from offset 0x01f1 of kernel image on should be loaded into struct
  636. boot_params and examined. The end of setup header can be calculated as
  637. follow:
  638. 0x0202 + byte value at offset 0x0201
  639. In addition to read/modify/write the setup header of the struct
  640. boot_params as that of 16-bit boot protocol, the boot loader should
  641. also fill the additional fields of the struct boot_params as that
  642. described in zero-page.txt.
  643. After setupping the struct boot_params, the boot loader can load the
  644. 32/64-bit kernel in the same way as that of 16-bit boot protocol.
  645. In 32-bit boot protocol, the kernel is started by jumping to the
  646. 32-bit kernel entry point, which is the start address of loaded
  647. 32/64-bit kernel.
  648. At entry, the CPU must be in 32-bit protected mode with paging
  649. disabled; a GDT must be loaded with the descriptors for selectors
  650. __BOOT_CS(0x10) and __BOOT_DS(0x18); both descriptors must be 4G flat
  651. segment; __BOOS_CS must have execute/read permission, and __BOOT_DS
  652. must have read/write permission; CS must be __BOOT_CS and DS, ES, SS
  653. must be __BOOT_DS; interrupt must be disabled; %esi must hold the base
  654. address of the struct boot_params; %ebp, %edi and %ebx must be zero.