1.\" Copyright (c) 1993 Paul Kranenburg 2.\" All rights reserved. 3.\" 4.\" Redistribution and use in source and binary forms, with or without 5.\" modification, are permitted provided that the following conditions 6.\" are met: 7.\" 1. Redistributions of source code must retain the above copyright 8.\" notice, this list of conditions and the following disclaimer. 9.\" 2. Redistributions in binary form must reproduce the above copyright 10.\" notice, this list of conditions and the following disclaimer in the 11.\" documentation and/or other materials provided with the distribution. 12.\" 3. All advertising materials mentioning features or use of this software 13.\" must display the following acknowledgement: 14.\" This product includes software developed by Paul Kranenburg. 15.\" 3. The name of the author may not be used to endorse or promote products 16.\" derived from this software without specific prior written permission 17.\" 18.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR 19.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 20.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 21.\" IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, 22.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 23.\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 24.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 25.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 26.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 27.\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 28.\" 29.\" $FreeBSD$ 30.\" 31.Dd October 23, 1993 32.Dt LINK 5 33.Os 34.Sh NAME 35.Nm link 36.Nd dynamic loader and link editor interface 37.Sh SYNOPSIS 38.Fd #include <sys/types.h> 39.Fd #include <nlist.h> 40.Fd #include <link.h> 41.Sh DESCRIPTION 42The include file 43.Aq Pa link.h 44declares several structures that are present in dynamically linked 45programs and libraries. 46The structures define the interface between several components of the 47link-editor and loader mechanism. The layout of a number of these 48structures within the binaries resembles the a.out format in many places 49as it serves such similar functions as symbol definitions (including the 50accompanying string table) and relocation records needed to resolve 51references to external entities. It also records a number of data structures 52unique to the dynamic loading and linking process. These include references 53to other objects that are required to complete the link-editing process and 54indirection tables to facilitate 55.Em Position Independent Code 56(PIC for short) to improve sharing of code pages among different processes. 57The collection of data structures described here will be referred to as the 58.Em Run-time Relocation Section (RRS) 59and is embedded in the standard text and data segments of the dynamically 60linked program or shared object image as the existing 61.Xr a.out 5 62format offers no room for it elsewhere. 63.Pp 64Several utilities cooperate to ensure that the task of getting a program 65ready to run can complete successfully in a way that optimizes the use 66of system resources. The compiler emits PIC code from which shared libraries 67can be built by 68.Xr ld 1 . 69The compiler also includes size information of any initialized data items 70through the .size assembler directive. PIC code differs from conventional code 71in that it accesses data variables through an indirection table, the 72Global Offset Table, by convention accessible by the reserved name 73.Em _GLOBAL_OFFSET_TABLE_. 74The exact mechanism used for this is machine dependent, usually a machine 75register is reserved for the purpose. The rational behind this construct 76is to generate code that is independent of the actual load address. Only 77the values contained in the Global Offset Table may need updating at run-time 78depending on the load addresses of the various shared objects in the address 79space. 80.Pp 81Likewise, procedure calls to globally defined functions are redirected through 82the Procedure Linkage Table (PLT) residing in the data segment of the core 83image. Again, this is done to avoid run-time modifications to the text segment. 84.Pp 85The linker-editor allocates the Global Offset Table and Procedure Linkage Table 86when combining PIC object files into an image suitable for mapping into the 87process address space. It also collects all symbols that may be needed by the 88run-time link-editor and stores these along with the image's text and data bits. 89Another reserved symbol, 90.Em _DYNAMIC 91is used to indicate the presence of the run-time linker structures. Whenever 92_DYNAMIC is relocated to 0, there is no need to invoke the run-time 93link-editor. If this symbol is non-zero, it points at a data structure from 94which the location of the necessary relocation- and symbol information can 95be derived. This is most notably used by the start-up module, 96.Em crt0. 97The _DYNAMIC structure is conventionally located at the start of the data 98segment of the image to which it pertains. 99.Pp 100.Sh DATA STRUCTURES 101The data structures supporting dynamic linking and run-time relocation 102reside both in the text and data segments of the image they apply to. 103The text segments contain read-only data such as symbols descriptions and 104names, while the data segments contain the tables that need to be modified by 105during the relocation process. 106.Pp 107The _DYNAMIC symbol references a 108.Fa _dynamic 109structure: 110.Bd -literal -offset indent 111struct _dynamic { 112 int d_version; 113 struct so_debug *d_debug; 114 union { 115 struct section_dispatch_table *d_sdt; 116 } d_un; 117 struct ld_entry *d_entry; 118}; 119.Ed 120.Bl -tag -width d_version 121.It Fa d_version 122This field provides for different versions of the dynamic linking 123implementation. The current version numbers understood by 124.Xr ld 1 125and 126.Xr ld.so 1 127are 128.Em LD_VERSION_SUN (3), 129which is used by the SunOS 4.x releases, and 130.Em LD_VERSION_BSD (8), 131which is currently in use by FreeBSD since release 1.1. 132.It Fa d_un 133Refers to a 134.Em d_version 135dependent data structure. 136.It Fa so_debug 137this field provides debuggers with a hook to access symbol tables of shared 138objects loaded as a result of the actions of the run-time link-editor. 139.El 140.Pp 141The 142.Fa section_dispatch_table 143structure is the main 144.Dq dispatcher 145table, containing offsets into the image's segments where various symbol 146and relocation information is located. 147.Bd -literal -offset indent 148struct section_dispatch_table { 149 struct so_map *sdt_loaded; 150 long sdt_sods; 151 long sdt_filler1; 152 long sdt_got; 153 long sdt_plt; 154 long sdt_rel; 155 long sdt_hash; 156 long sdt_nzlist; 157 long sdt_filler2; 158 long sdt_buckets; 159 long sdt_strings; 160 long sdt_str_sz; 161 long sdt_text_sz; 162 long sdt_plt_sz; 163}; 164.Ed 165.Pp 166.Bl -tag -width sdt_filler1 167.It Fa sdt_loaded 168A pointer to the first link map loaded (see below). This field is set by 169.Nm ld.so 170.It Fa sdt_sods 171The start of a (linked) list of shared object descriptors needed by 172.Em this 173object. 174.It Fa sdt_filler1 175Deprecated (used by SunOS to specify library search rules). 176.It Fa sdt_got 177The location of the Global Offset Table within this image. 178.It Fa sdt_plt 179The location of the Procedure Linkage Table within this image. 180.It Fa sdt_rel 181The location of an array of 182.Fa relocation_info 183structures 184.Po 185see 186.Xr a.out 5 187.Pc 188specifying run-time relocations. 189.It Fa sdt_hash 190The location of the hash table for fast symbol lookup in this object's 191symbol table. 192.It Fa sdt_nzlist 193The location of the symbol table. 194.It Fa sdt_filler2 195Currently unused. 196.It Fa sdt_buckets 197The number of buckets in 198.Fa sdt_hash 199.It Fa sdt_strings 200The location of the symbol string table that goes with 201.Fa sdt_nzlist. 202.It Fa sdt_str_sz 203The size of the string table. 204.It Fa sdt_text_sz 205The size of the object's text segment. 206.It Fa sdt_plt_sz 207The size of the Procedure Linkage Table. 208.El 209.Pp 210A 211.Fa sod 212structure describes a shared object that is needed 213to complete the link edit process of the object containing it. 214A list of such objects 215.Po 216chained through 217.Fa sod_next 218.Pc 219is pointed at 220by the 221.Fa sdt_sods 222in the section_dispatch_table structure. 223.Bd -literal -offset indent 224struct sod { 225 long sod_name; 226 u_int sod_library : 1, 227 sod_reserved : 31; 228 short sod_major; 229 short sod_minor; 230 long sod_next; 231}; 232.Ed 233.Pp 234.Bl -tag -width sod_library 235.It Fa sod_name 236The offset in the text segment of a string describing this link object. 237.It Fa sod_library 238If set, 239.Fa sod_name 240specifies a library that is to be searched for by 241.Nm ld.so . 242The path name 243is obtained by searching a set of directories 244.Po 245see also 246.Xr ldconfig 8 247.Pc 248for a shared object matching 249.Em lib\&<sod_name>\&.so.n.m. 250If not set, 251.Fa sod_name 252should point at a full path name for the desired shared object. 253.It Fa sod_major 254Specifies the major version number of the shared object to load. 255.It Fa sod_minor 256Specifies the prefered minor version number of the shared object to load. 257.El 258.Pp 259The run-time link-editor maintains a list of structures called 260.Em link maps 261to keep track of all shared objects loaded into a process' address space. 262These structures are only used at run-time and do not occur within 263the text or data segment of an executable or shared library. 264.Bd -literal -offset indent 265struct so_map { 266 caddr_t som_addr; 267 char *som_path; 268 struct so_map *som_next; 269 struct sod *som_sod; 270 caddr_t som_sodbase; 271 u_int som_write : 1; 272 struct _dynamic *som_dynamic; 273 caddr_t som_spd; 274}; 275.Ed 276.Bl -tag -width som_dynamic 277.It Fa som_addr 278The address at which the shared object associated with this link map has 279been loaded. 280.It Fa som_path 281The full path name of the loaded object. 282.It Fa som_next 283Pointer to the next link map. 284.It Fa som_sod 285The 286.Fa sod 287structure that was responsible for loading this shared object. 288.It Fa som_sodbase 289Tossed in later versions the run-time linker. 290.It Fa som_write 291Set if (some portion of) this object's text segment is currently writable. 292.It Fa som_dynamic 293Pointer to this object's 294.Fa _dynamic 295structure. 296.It Fa som_spd 297Hook for attaching private data maintained by the run-time link-editor. 298.El 299.Pp 300Symbol description with size. This is simply an 301.Fa nlist 302structure with one field 303.Pq Fa nz_size 304added. Used to convey size information on items in the data segment 305of shared objects. An array of these lives in the shared object's 306text segment and is addressed by the 307.Fa sdt_nzlist 308field of 309.Fa section_dispatch_table. 310.Bd -literal -offset indent 311struct nzlist { 312 struct nlist nlist; 313 u_long nz_size; 314#define nz_un nlist.n_un 315#define nz_strx nlist.n_un.n_strx 316#define nz_name nlist.n_un.n_name 317#define nz_type nlist.n_type 318#define nz_value nlist.n_value 319#define nz_desc nlist.n_desc 320#define nz_other nlist.n_other 321}; 322.Ed 323.Bl -tag -width nz_size 324.It Fa nlist 325.Po 326see 327.Xr nlist 3 328.Pc . 329.It Fa nz_size 330The size of the data represented by this symbol. 331.El 332.Pp 333A hash table is included within the text segment of shared object to 334to facilitate quick lookup of symbols during run-time link-editing. 335The 336.Fa sdt_hash 337field of the 338.Fa section_dispatch_table 339structure points at an array of 340.Fa rrs_hash 341structures: 342.Bd -literal -offset indent 343struct rrs_hash { 344 int rh_symbolnum; /* symbol number */ 345 int rh_next; /* next hash entry */ 346}; 347.Ed 348.Pp 349.Bl -tag -width rh_symbolnum 350.It Fa rh_symbolnum 351The index of the symbol in the shared object's symbol table (as given by the 352.Fa ld_symbols 353field). 354.It Fa rh_next 355In case of collisions, this field is the offset of the next entry in this 356hash table bucket. It is zero for the last bucket element. 357.El 358The 359.Fa rt_symbol 360structure is used to keep track of run-time allocated commons 361and data items copied from shared objects. These items are kept on linked list 362and is exported through the 363.Fa dd_cc 364field in the 365.Fa so_debug 366structure (see below) for use by debuggers. 367.Bd -literal -offset indent 368struct rt_symbol { 369 struct nzlist *rt_sp; 370 struct rt_symbol *rt_next; 371 struct rt_symbol *rt_link; 372 caddr_t rt_srcaddr; 373 struct so_map *rt_smp; 374}; 375.Ed 376.Pp 377.Bl -tag -width rt_scraddr 378.It Fa rt_sp 379The symbol description. 380.It Fa rt_next 381Virtual address of next rt_symbol. 382.It Fa rt_link 383Next in hash bucket. Used by internally by 384.Nm ld.so . 385.It Fa rt_srcaddr 386Location of the source of initialized data within a shared object. 387.It Fa rt_smp 388The shared object which is the original source of the data that this 389run-time symbol describes. 390.El 391.Pp 392The 393.Fa so_debug 394structure is used by debuggers to gain knowledge of any shared objects 395that have been loaded in the process's address space as a result of run-time 396link-editing. Since the run-time link-editor runs as a part of process 397initialization, a debugger that wishes to access symbols from shared objects 398can only do so after the link-editor has been called from crt0. 399A dynamically linked binary contains a 400.Fa so_debug 401structure which can be located by means of the 402.Fa d_debug 403field in 404.Fa _dynamic . 405.Bd -literal -offset indent 406struct so_debug { 407 int dd_version; 408 int dd_in_debugger; 409 int dd_sym_loaded; 410 char *dd_bpt_addr; 411 int dd_bpt_shadow; 412 struct rt_symbol *dd_cc; 413}; 414.Ed 415.Pp 416.Bl -tag -width dd_in_debugger 417.It Fa dd_version 418Version number of this interface. 419.It Fa dd_in_debugger 420Set by the debugger to indicate to the run-time linker that the program is 421run under control of a debugger. 422.It Fa dd_sym_loaded 423Set by the run-time linker whenever it adds symbols by loading shared objects. 424.It Fa dd_bpt_addr 425The address were a breakpoint will be set by the run-time linker to 426divert control to the debugger. This address is determined by the start-up 427module, 428.Em crt0.o, 429to be some convenient place before the call to _main. 430.It Fa dd_bpt_shadow 431Contains the original instruction that was at 432.Fa dd_bpt_addr . 433The debugger is expected to put this instruction back before continuing the 434program. 435.It Fa dd_cc 436A pointer to the linked list of run-time allocated symbols that the debugger 437may be interested in. 438.El 439.Pp 440The 441.Em ld_entry 442structure defines a set of service routines within 443.Nm ld.so . 444.\" See 445.\" .Xr libdl.a 446.\" for more information. 447.Bd -literal -offset indent 448struct ld_entry { 449 void *(*dlopen)(char *, int); 450 int (*dlclose)(void *); 451 void *(*dlsym)(void *, char *); 452 char *(*dlerror)(void); 453}; 454.Ed 455 456The 457.Fa crt_ldso 458structure defines the interface between the start-up code in crt0 and 459.Nm ld.so . 460.Bd -literal -offset indent 461struct crt_ldso { 462 int crt_ba; 463 int crt_dzfd; 464 int crt_ldfd; 465 struct _dynamic *crt_dp; 466 char **crt_ep; 467 caddr_t crt_bp; 468 char *crt_prog; 469 char *crt_ldso; 470 struct ld_entry *crt_ldentry; 471}; 472#define CRT_VERSION_SUN 1 473#define CRT_VERSION_BSD_2 2 474#define CRT_VERSION_BSD_3 3 475#define CRT_VERSION_BSD_4 4 476.Ed 477.Bl -tag -width crt_dzfd 478.It Fa crt_ba 479The virtual address at which 480.Nm ld.so 481was loaded by crt0. 482.It Fa crt_dzfd 483On SunOS systems, this field contains an open file descriptor to 484.Dq Pa /dev/zero 485used to get demand paged zeroed pages. On FreeBSD systems it contains -1. 486.It Fa crt_ldfd 487Contains an open file descriptor that was used by crt0 to load 488.Nm ld.so . 489.It Fa crt_dp 490A pointer to main's 491.Fa _dynamic 492structure. 493.It Fa crt_ep 494A pointer to the environment strings. 495.It Fa crt_bp 496The address at which a breakpoint will be placed by the run-time linker 497if the main program is run by a debugger. 498See 499.Fa so_debug 500.It Fa crt_prog 501The name of the main program as determined by crt0 (CRT_VERSION_BSD3 only). 502.It Fa crt_ldso 503The path of the run-time linker as mapped by crt0 (CRT_VERSION_BSD4 only). 504.El 505.Pp 506The 507.Fa hints_header 508and 509.Fa hints_bucket 510structures define the layout of the library hints, normally found in 511.Dq Pa /var/run/ld.so.hints , 512which is used by 513.Nm ld.so 514to quickly locate the shared object images in the 515filesystem. 516The organization of the hints file is not unlike that of an 517.Dq a.out 518object file, in that it contains a header determining the offset and size 519of a table of fixed sized hash buckets and a common string pool. 520.Bd -literal -offset indent 521struct hints_header { 522 long hh_magic; 523#define HH_MAGIC 011421044151 524 long hh_version; 525#define LD_HINTS_VERSION_1 1 526 long hh_hashtab; 527 long hh_nbucket; 528 long hh_strtab; 529 long hh_strtab_sz; 530 long hh_ehints; 531}; 532.Ed 533.Bl -tag -width hh_strtab_sz 534.It Fa hh_magic 535Hints file magic number. 536.It Fa hh_version 537Interface version number. 538.It Fa hh_hashtab 539Offset of hash table. 540.It Fa hh_strtab 541Offset of string table. 542.It Fa hh_strtab_sz 543Size of strings. 544.It Fa hh_ehints 545Maximum usable offset in hints file. 546.El 547.Pp 548.Bd -literal -offset indent 549/* 550 * Hash table element in hints file. 551 */ 552struct hints_bucket { 553 int hi_namex; 554 int hi_pathx; 555 int hi_dewey[MAXDEWEY]; 556 int hi_ndewey; 557#define hi_major hi_dewey[0] 558#define hi_minor hi_dewey[1] 559 int hi_next; 560}; 561.Ed 562.Bl -tag -width hi_ndewey 563.It Fa hi_namex 564Index of the string identifying the library. 565.It Fa hi_pathx 566Index of the string representing the full path name of the library. 567.It Fa hi_dewey 568The version numbers of the shared library. 569.It Fa hi_ndewey 570The number of valid entries in 571.Fa hi_dewey . 572.It Fa hi_next 573Next bucket in case of hashing collisions. 574.El 575.Sh CAVEATS 576Only the (GNU) C compiler currently supports the creation of shared libraries. 577Other programming languages can not be used. 578