ack.out.5 10 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317
  1. .TH "ACK.OUT" 5 "$Revision$"
  2. .ad
  3. .SH NAME
  4. ack.out\ \-\ ACK-assembler and link editor output
  5. .SH SYNOPSIS
  6. .B #include <out.h>
  7. .SH DESCRIPTION
  8. This manual page discusses the format of object files, as generated by ACK
  9. assemblers and the link editor LED.
  10. The format is designed to be compact, machine independent, and
  11. portable from one machine to another,
  12. so that an object file can be produced on one machine, and
  13. further processed on another.
  14. .ta \w'#define x'u +\w'XXXXXXXX'u +\w'XXXXXXXXXXX'u
  15. .PP
  16. In the following discussion, some structures are defined using
  17. \fBlong\fR and \fBshort\fR as type indicators.
  18. It is assumed that the size of a short is 2 bytes (chars) and that the
  19. size of a long is 4 bytes.
  20. However, these types
  21. have a machine dependent byte and word order.
  22. Therefore, a machine independent representation is chosen for the
  23. object format:
  24. a long consists of two shorts, of which the least significant one
  25. comes first, and a short consists of two bytes, of which the
  26. least significant one comes first.
  27. There is no alignment between various parts and structures in the object
  28. file.
  29. .PP
  30. In general, an object file consists of the following parts:
  31. .PP
  32. .nf
  33. \- a file header
  34. \- a number of section headers
  35. \- the sections themselves
  36. \- a number of relocation structures
  37. \- a symbol table
  38. \- a string area containing the names from the symbol table
  39. .fi
  40. .PP
  41. .B The header.
  42. .br
  43. The header of an object file has the following structure:
  44. .PP
  45. .nf
  46. struct outhead {
  47. unsigned short oh_magic; /* magic number */
  48. unsigned short oh_stamp; /* version stamp */
  49. unsigned short oh_flags; /* several format flags */
  50. unsigned short oh_nsect; /* number of outsect structures */
  51. unsigned short oh_nrelo; /* number of outrelo structures */
  52. unsigned short oh_nname; /* number of outname structures */
  53. long oh_nemit; /* length of sections */
  54. long oh_nchar; /* size of string area */
  55. };
  56. .fi
  57. .PP
  58. #define HF_LINK 0x0004 /* unresolved references left */
  59. .PP
  60. The fields of this structure have the following purpose:
  61. .nr x \w'oh_magic\ \ \ 'u
  62. .IP oh_magic \nxu
  63. A magic number, indicating that this is an object file.
  64. .IP oh_stamp \nxu
  65. A version stamp, used to detect obsolete versions of object files.
  66. .IP oh_flags \nxu
  67. Currently only used for the HF_LINK flag. When this flag is set, the
  68. object file contains unresolved references.
  69. .IP oh_nsect \nxu
  70. The number of sections and section description structures, later on
  71. referred to as \fIoutsect\fR structures.
  72. Usually, there are only a few sections, f.i. a TEXT section,
  73. a ROM section, a DATA section and a BSS section.
  74. Notice that neither the assemblers nor LED know more about them than their
  75. names.
  76. .IP oh_nrelo \nxu
  77. The number of relocation structures, later on referred to as \fIoutrelo\fR
  78. structures.
  79. .IP oh_nname \nxu
  80. The number of symbol table structures, later on referred to as \fIoutname\fR
  81. structures.
  82. .IP oh_nemit \nxu
  83. The total number of bytes in this object file used for the sections themselves.
  84. This field is used to find the relocation and symbol table structures fast.
  85. .IP oh_nchar \nxu
  86. The size of the string area (the number of bytes).
  87. .PP
  88. .B The section descriptions.
  89. .br
  90. The next part of an object file contains the outsect-structures.
  91. An outsect structure has the following layout:
  92. .PP
  93. .nf
  94. struct outsect {
  95. long os_base; /* start address in machine */
  96. long os_size; /* section size in machine */
  97. long os_foff; /* start address in file */
  98. long os_flen; /* section size in file */
  99. long os_lign; /* section alignment */
  100. };
  101. .fi
  102. .PP
  103. The fields in this structure have the following purpose:
  104. .IP os_base \nxu
  105. The start address of this section in the target machine.
  106. This address is determined by LED,
  107. when producing a non-relocatable object file.
  108. It is ignored for relocatable object files.
  109. .IP os_size \nxu
  110. The size of this section on the target machine.
  111. .IP os_foff \nxu
  112. The start address of this section in this file.
  113. .IP os_flen \nxu
  114. The size of this section in this file.
  115. This field does not have to have
  116. the same value as the \fIos_size\fR field!
  117. For instance, an uninitialized
  118. data section probably has \fIos_flen\fR set to 0.
  119. Notice that
  120. the \fIoh_nemit\fR field of the header contains
  121. the sum of all the \fIos_flen\fR fields.
  122. .IP os_lign \nxu
  123. The alignment requirement for this section. The requirement is that
  124. the loader must leave
  125. .IP "" \nxu
  126. \ \ \ \ \ \ \ \fIos_base\fR \fBmod\fR \fIos_lign\fR = 0
  127. .IP "" \nxu
  128. in tact.
  129. .PP
  130. .B The sections.
  131. .br
  132. The next part of an object file contains the sections themselves.
  133. Usually, the LED program places the sections right behind one another in the
  134. target machine, taking the
  135. alignment requirements into account. However, the user is allowed to give
  136. the start addresses of each section. But if the user gave a start address for
  137. say section 2, but not for section 3, section 3 will be put
  138. right behind section 2.
  139. .PP
  140. .B The relocation structures.
  141. .br
  142. Relocation information is information that allows a program like LED
  143. to combine several object files and produce an executable binary
  144. if there are no unresolved references.
  145. If relocation information is present, it amounts to 8 bytes per
  146. relocatable datum. The information has the following structure:
  147. .PP
  148. .nf
  149. struct outrelo {
  150. char or_type; /* type of reference */
  151. char or_sect; /* referencing section */
  152. unsigned short or_nami; /* referenced symbol index */
  153. long or_addr; /* referencing address */
  154. };
  155. .fi
  156. .PP
  157. .nf
  158. /*
  159. * relocation type bits
  160. */
  161. #define RELSZ 0x07 /* relocation length */
  162. #define RELO1 0x01 /* 1 byte */
  163. #define RELO2 0x02 /* 2 bytes */
  164. #define RELO4 0x04 /* 4 bytes */
  165. #define RELPC 0x08 /* pc relative */
  166. #define RELBR 0x10 /* High order byte lowest address. */
  167. #define RELWR 0x20 /* High order word lowest address. */
  168. .fi
  169. .PP
  170. .nf
  171. /*
  172. * section type bits and fields
  173. */
  174. #define S_TYP 0x007F /* undefined, absolute or relative */
  175. #define S_EXT 0x0080 /* external flag */
  176. #define S_ETC 0x7F00 /* for symbolic debug, bypassing 'as' */
  177. .fi
  178. .PP
  179. .nf
  180. /*
  181. * S_TYP field values
  182. */
  183. #define S_UND 0x0000 /* undefined item */
  184. #define S_ABS 0x0001 /* absolute item */
  185. #define S_MIN 0x0002 /* first user section */
  186. #define S_MAX (S_TYP-1) /* last user section */
  187. #define S_CRS S_TYP /* reference to other namelist item */
  188. .fi
  189. .PP
  190. The fields of this structure have the following purpose:
  191. .IP or_type \nxu
  192. Contains several flags: One of RELO1, RELO2 and RELO4 is set, indicating the
  193. size of the relocatable datum, RELPC is set when the datum is
  194. relocated pc relative, RELBR and RELWR indicate byte and word order of
  195. the relocatable datum. RELBR and RELWR are needed here. It is not sufficient
  196. to have flags for them in the header of the object file, because some
  197. machines (NS 32016) use several of the possible combinations in their
  198. instruction encoding.
  199. .IP or_sect \nxu
  200. Contains the section number of the referenc\fIing\fR section. This is a number
  201. that lies between S_MIN and S_MAX. The section indicated with number S_MIN
  202. is the first section in the sections-section, etc.
  203. .IP or_addr \nxu
  204. Contains the address of the relocatable datum, in the form of an
  205. offset from the base of the section indicated in the \fIor_sect\fR field.
  206. .IP or_nami \nxu
  207. Usually contains the index of the referenced symbol in the symbol table,
  208. starting at 0.
  209. In this case, the reference is to an undefined external symbol, a common
  210. symbol, or a section name. The relocatable datum then contains
  211. an offset from the indicated symbol or the start of the indicated section.
  212. It may, however, also have the same value as
  213. the \fIoh_nname\fR field of the header. In this case the relocatable datum
  214. is an absolute number, and the datum is relocated pc relative.
  215. The relocatable datum must then be relocated with respect to the
  216. base address of its section.
  217. .PP
  218. .B The symbol table.
  219. .br
  220. This table contains definitions of symbols. It is referred to by
  221. outrelo-structures, and can be used by debuggers.
  222. Entries in this table have the following structure:
  223. .PP
  224. .nf
  225. struct outname {
  226. union {
  227. char *on_ptr; /* symbol name (in core) */
  228. long on_off; /* symbol name (in file) */
  229. } on_u;
  230. #define on_mptr on_u.on_ptr
  231. #define on_foff on_u.on_off
  232. unsigned short on_type; /* symbol type */
  233. unsigned short on_desc; /* debug info */
  234. long on_valu; /* symbol value */
  235. };
  236. .fi
  237. .PP
  238. .nf
  239. /*
  240. * S_ETC field values
  241. */
  242. #define S_SCT 0x0100 /* section names */
  243. #define S_LIN 0x0200 /* hll source line item */
  244. #define S_FIL 0x0300 /* hll source file item */
  245. #define S_MOD 0x0400 /* ass source file item */
  246. #define S_COM 0x1000 /* Common name */
  247. .fi
  248. .PP
  249. The members of this structure have the following purpose:
  250. .IP on_foff \nxu
  251. Contains the offset of the name from the beginning of the file. The name
  252. extends from the offset to the next null byte.
  253. .IP on_type \nxu
  254. The S_TYP field of this member contains the section number of the symbol.
  255. Here, this number may be S_ABS for an absolute item, or S_UND, for an
  256. undefined item. The S_EXT flag is set in this member if the symbol is external.
  257. The S_ETC field has the following flags:
  258. S_SCT is set if the symbol represents a section name,
  259. S_COM is set if the symbol represents a common name,
  260. S_LIN is set if the symbol refers to a high level language source line item,
  261. S_FIL is set if the symbol refers to a high level language source file item,
  262. and S_MOD is set if the symbol refers to an assembler source file item.
  263. .IP on_desc \nxu
  264. Currently not used.
  265. .IP on_valu \nxu
  266. Is not used if the symbol refers to an undefined item. For absolute items
  267. it contains the value, for common names it contains the size, and
  268. for anything else it contains the offset from the beginning of the section.
  269. In a fully linked binary, the beginning of the section is added.
  270. .PP
  271. .B The string area.
  272. .br
  273. The last part of an object file contains the name list. This is just a
  274. sequence of null-terminated strings.
  275. .PP
  276. The relocation information, the symbol table, and the name list do not
  277. have to be present, but then of course we do not have a relocatable
  278. object file.
  279. .PP
  280. .B Miscellaneous defines
  281. .br
  282. The following miscellaneous defines might come in handy when reading
  283. object files:
  284. .PP
  285. .nf
  286. /*
  287. * structure format strings
  288. */
  289. #define SF_HEAD "22222244"
  290. #define SF_SECT "44444"
  291. #define SF_RELO "1124"
  292. #define SF_NAME "4224"
  293. .fi
  294. .PP
  295. .nf
  296. /*
  297. * structure sizes (bytes in file; add digits in SF_*)
  298. */
  299. #define SZ_HEAD 20
  300. #define SZ_SECT 20
  301. #define SZ_RELO 8
  302. #define SZ_NAME 12
  303. .fi
  304. .PP
  305. .nf
  306. /*
  307. * file access macros
  308. */
  309. #define BADMAGIC(x) ((x).oh_magic!=O_MAGIC)
  310. #define OFF_SECT(x) SZ_HEAD
  311. #define OFF_EMIT(x) (OFF_SECT(x) + ((long)(x).oh_nsect * SZ_SECT))
  312. #define OFF_RELO(x) (OFF_EMIT(x) + (x).oh_nemit)
  313. #define OFF_NAME(x) (OFF_RELO(x) + ((long)(x).oh_nrelo * SZ_RELO))
  314. #define OFF_CHAR(x) (OFF_NAME(x) + ((long)(x).oh_nname * SZ_NAME))
  315. .fi
  316. .SH "SEE ALSO"
  317. led(6), object(3)