xref: /freebsd/share/man/man5/link.5 (revision 6780ab54325a71e7e70112b11657973edde8655e)
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.In sys/types.h
39.In nlist.h
40.In 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.
48The layout of a number of these
49structures within the binaries resembles the a.out format in many places
50as it serves such similar functions as symbol definitions (including the
51accompanying string table) and relocation records needed to resolve
52references to external entities.
53It also records a number of data structures
54unique to the dynamic loading and linking process.
55These include references
56to other objects that are required to complete the link-editing process and
57indirection tables to facilitate
58.Em Position Independent Code
59(PIC for short) to improve sharing of code pages among different processes.
60The collection of data structures described here will be referred to as the
61.Em Run-time Relocation Section (RRS)
62and is embedded in the standard text and data segments of the dynamically
63linked program or shared object image as the existing
64.Xr a.out 5
65format offers no room for it elsewhere.
66.Pp
67Several utilities cooperate to ensure that the task of getting a program
68ready to run can complete successfully in a way that optimizes the use
69of system resources.
70The compiler emits PIC code from which shared libraries
71can be built by
72.Xr ld 1 .
73The compiler also includes size information of any initialized data items
74through the .size assembler directive.
75PIC code differs from conventional code
76in that it accesses data variables through an indirection table, the
77Global Offset Table, by convention accessible by the reserved name
78.Dv _GLOBAL_OFFSET_TABLE_ .
79The exact mechanism used for this is machine dependent, usually a machine
80register is reserved for the purpose.
81The rational behind this construct
82is to generate code that is independent of the actual load address.
83Only
84the values contained in the Global Offset Table may need updating at run-time
85depending on the load addresses of the various shared objects in the address
86space.
87.Pp
88Likewise, procedure calls to globally defined functions are redirected through
89the Procedure Linkage Table (PLT) residing in the data segment of the core
90image.
91Again, this is done to avoid run-time modifications to the text segment.
92.Pp
93The linker-editor allocates the Global Offset Table and Procedure Linkage Table
94when combining PIC object files into an image suitable for mapping into the
95process address space.
96It also collects all symbols that may be needed by the
97run-time link-editor and stores these along with the image's text and data bits.
98Another reserved symbol,
99.Em _DYNAMIC
100is used to indicate the presence of the run-time linker structures.
101Whenever
102_DYNAMIC is relocated to 0, there is no need to invoke the run-time
103link-editor.
104If this symbol is non-zero, it points at a data structure from
105which the location of the necessary relocation- and symbol information can
106be derived.
107This is most notably used by the start-up module,
108.Em crt0 .
109The _DYNAMIC structure is conventionally located at the start of the data
110segment of the image to which it pertains.
111.Sh DATA STRUCTURES
112The data structures supporting dynamic linking and run-time relocation
113reside both in the text and data segments of the image they apply to.
114The text segments contain read-only data such as symbols descriptions and
115names, while the data segments contain the tables that need to be modified by
116during the relocation process.
117.Pp
118The _DYNAMIC symbol references a
119.Fa _dynamic
120structure:
121.Bd -literal -offset indent
122struct	_dynamic {
123	int	d_version;
124	struct 	so_debug *d_debug;
125	union {
126		struct section_dispatch_table *d_sdt;
127	} d_un;
128	struct  ld_entry *d_entry;
129};
130.Ed
131.Bl -tag -width d_version
132.It Fa d_version
133This field provides for different versions of the dynamic linking
134implementation.
135The current version numbers understood by
136.Xr ld 1
137and
138.Xr ld.so 1
139are
140.Em LD_VERSION_SUN (3) ,
141which is used by the
142.Tn SunOS
1434.x releases, and
144.Em LD_VERSION_BSD (8) ,
145which has been in use since
146.Fx 1.1 .
147.It Fa d_un
148Refers to a
149.Em d_version
150dependent data structure.
151.It Fa so_debug
152this field provides debuggers with a hook to access symbol tables of shared
153objects loaded as a result of the actions of the run-time link-editor.
154.El
155.Pp
156The
157.Fa section_dispatch_table
158structure is the main
159.Dq dispatcher
160table, containing offsets into the image's segments where various symbol
161and relocation information is located.
162.Bd -literal -offset indent
163struct section_dispatch_table {
164	struct	so_map *sdt_loaded;
165	long	sdt_sods;
166	long	sdt_filler1;
167	long	sdt_got;
168	long	sdt_plt;
169	long	sdt_rel;
170	long	sdt_hash;
171	long	sdt_nzlist;
172	long	sdt_filler2;
173	long	sdt_buckets;
174	long	sdt_strings;
175	long	sdt_str_sz;
176	long	sdt_text_sz;
177	long	sdt_plt_sz;
178};
179.Ed
180.Pp
181.Bl -tag -width sdt_filler1
182.It Fa sdt_loaded
183A pointer to the first link map loaded (see below). This field is set by
184.Nm ld.so
185.It Fa sdt_sods
186The start of a (linked) list of shared object descriptors needed by
187.Em this
188object.
189.It Fa sdt_filler1
190Deprecated (used by SunOS to specify library search rules).
191.It Fa sdt_got
192The location of the Global Offset Table within this image.
193.It Fa sdt_plt
194The location of the Procedure Linkage Table within this image.
195.It Fa sdt_rel
196The location of an array of
197.Fa relocation_info
198structures
199(see
200.Xr a.out 5 )
201specifying run-time relocations.
202.It Fa sdt_hash
203The location of the hash table for fast symbol lookup in this object's
204symbol table.
205.It Fa sdt_nzlist
206The location of the symbol table.
207.It Fa sdt_filler2
208Currently unused.
209.It Fa sdt_buckets
210The number of buckets in
211.Fa sdt_hash
212.It Fa sdt_strings
213The location of the symbol string table that goes with
214.Fa sdt_nzlist .
215.It Fa sdt_str_sz
216The size of the string table.
217.It Fa sdt_text_sz
218The size of the object's text segment.
219.It Fa sdt_plt_sz
220The size of the Procedure Linkage Table.
221.El
222.Pp
223A
224.Fa sod
225structure describes a shared object that is needed
226to complete the link edit process of the object containing it.
227A list of such objects
228(chained through
229.Fa sod_next )
230is pointed at
231by the
232.Fa sdt_sods
233in the section_dispatch_table structure.
234.Bd -literal -offset indent
235struct sod {
236	long	sod_name;
237	u_int	sod_library : 1,
238		sod_reserved : 31;
239	short	sod_major;
240	short	sod_minor;
241	long	sod_next;
242};
243.Ed
244.Pp
245.Bl -tag -width sod_library
246.It Fa sod_name
247The offset in the text segment of a string describing this link object.
248.It Fa sod_library
249If set,
250.Fa sod_name
251specifies a library that is to be searched for by
252.Nm ld.so .
253The path name
254is obtained by searching a set of directories
255(see also
256.Xr ldconfig 8 )
257for a shared object matching
258.Em lib\&<sod_name>\&.so.n.m .
259If not set,
260.Fa sod_name
261should point at a full path name for the desired shared object.
262.It Fa sod_major
263Specifies the major version number of the shared object to load.
264.It Fa sod_minor
265Specifies the preferred minor version number of the shared object to load.
266.El
267.Pp
268The run-time link-editor maintains a list of structures called
269.Em link maps
270to keep track of all shared objects loaded into a process' address space.
271These structures are only used at run-time and do not occur within
272the text or data segment of an executable or shared library.
273.Bd -literal -offset indent
274struct so_map {
275	caddr_t	som_addr;
276	char 	*som_path;
277	struct	so_map *som_next;
278	struct	sod *som_sod;
279	caddr_t som_sodbase;
280	u_int	som_write : 1;
281	struct	_dynamic *som_dynamic;
282	caddr_t	som_spd;
283};
284.Ed
285.Bl -tag -width som_dynamic
286.It Fa som_addr
287The address at which the shared object associated with this link map has
288been loaded.
289.It Fa som_path
290The full path name of the loaded object.
291.It Fa som_next
292Pointer to the next link map.
293.It Fa som_sod
294The
295.Fa sod
296structure that was responsible for loading this shared object.
297.It Fa som_sodbase
298Tossed in later versions the run-time linker.
299.It Fa som_write
300Set if (some portion of) this object's text segment is currently writable.
301.It Fa som_dynamic
302Pointer to this object's
303.Fa _dynamic
304structure.
305.It Fa som_spd
306Hook for attaching private data maintained by the run-time link-editor.
307.El
308.Pp
309Symbol description with size.
310This is simply an
311.Fa nlist
312structure with one field
313.Pq Fa nz_size
314added.
315Used to convey size information on items in the data segment
316of shared objects.
317An array of these lives in the shared object's
318text segment and is addressed by the
319.Fa sdt_nzlist
320field of
321.Fa section_dispatch_table .
322.Bd -literal -offset indent
323struct nzlist {
324	struct nlist	nlist;
325	u_long		nz_size;
326#define nz_un		nlist.n_un
327#define nz_strx		nlist.n_un.n_strx
328#define nz_name		nlist.n_un.n_name
329#define nz_type		nlist.n_type
330#define nz_value	nlist.n_value
331#define nz_desc		nlist.n_desc
332#define nz_other	nlist.n_other
333};
334.Ed
335.Bl -tag -width nz_size
336.It Fa nlist
337(see
338.Xr nlist 3 ) .
339.It Fa nz_size
340The size of the data represented by this symbol.
341.El
342.Pp
343A hash table is included within the text segment of shared object
344to facilitate quick lookup of symbols during run-time link-editing.
345The
346.Fa sdt_hash
347field of the
348.Fa section_dispatch_table
349structure points at an array of
350.Fa rrs_hash
351structures:
352.Bd -literal -offset indent
353struct rrs_hash {
354	int	rh_symbolnum;		/* symbol number */
355	int	rh_next;		/* next hash entry */
356};
357.Ed
358.Pp
359.Bl -tag -width rh_symbolnum
360.It Fa rh_symbolnum
361The index of the symbol in the shared object's symbol table (as given by the
362.Fa ld_symbols
363field).
364.It Fa rh_next
365In case of collisions, this field is the offset of the next entry in this
366hash table bucket.
367It is zero for the last bucket element.
368.El
369The
370.Fa rt_symbol
371structure is used to keep track of run-time allocated commons
372and data items copied from shared objects.
373These items are kept on linked list
374and is exported through the
375.Fa dd_cc
376field in the
377.Fa so_debug
378structure (see below) for use by debuggers.
379.Bd -literal -offset indent
380struct rt_symbol {
381	struct nzlist		*rt_sp;
382	struct rt_symbol	*rt_next;
383	struct rt_symbol	*rt_link;
384	caddr_t			rt_srcaddr;
385	struct so_map		*rt_smp;
386};
387.Ed
388.Pp
389.Bl -tag -width rt_scraddr
390.It Fa rt_sp
391The symbol description.
392.It Fa rt_next
393Virtual address of next rt_symbol.
394.It Fa rt_link
395Next in hash bucket.
396Used by internally by
397.Nm ld.so .
398.It Fa rt_srcaddr
399Location of the source of initialized data within a shared object.
400.It Fa rt_smp
401The shared object which is the original source of the data that this
402run-time symbol describes.
403.El
404.Pp
405The
406.Fa so_debug
407structure is used by debuggers to gain knowledge of any shared objects
408that have been loaded in the process's address space as a result of run-time
409link-editing.
410Since the run-time link-editor runs as a part of process
411initialization, a debugger that wishes to access symbols from shared objects
412can only do so after the link-editor has been called from crt0.
413A dynamically linked binary contains a
414.Fa so_debug
415structure which can be located by means of the
416.Fa d_debug
417field in
418.Fa _dynamic .
419.Bd -literal -offset indent
420struct 	so_debug {
421	int	dd_version;
422	int	dd_in_debugger;
423	int	dd_sym_loaded;
424	char    *dd_bpt_addr;
425	int	dd_bpt_shadow;
426	struct rt_symbol *dd_cc;
427};
428.Ed
429.Pp
430.Bl -tag -width dd_in_debugger
431.It Fa dd_version
432Version number of this interface.
433.It Fa dd_in_debugger
434Set by the debugger to indicate to the run-time linker that the program is
435run under control of a debugger.
436.It Fa dd_sym_loaded
437Set by the run-time linker whenever it adds symbols by loading shared objects.
438.It Fa dd_bpt_addr
439The address were a breakpoint will be set by the run-time linker to
440divert control to the debugger.
441This address is determined by the start-up
442module,
443.Pa crt0.o ,
444to be some convenient place before the call to _main.
445.It Fa dd_bpt_shadow
446Contains the original instruction that was at
447.Fa dd_bpt_addr .
448The debugger is expected to put this instruction back before continuing the
449program.
450.It Fa dd_cc
451A pointer to the linked list of run-time allocated symbols that the debugger
452may be interested in.
453.El
454.Pp
455The
456.Em ld_entry
457structure defines a set of service routines within
458.Nm ld.so .
459.\" See
460.\" .Xr libdl.a
461.\" for more information.
462.Bd -literal -offset indent
463struct ld_entry {
464	void	*(*dlopen)(char *, int);
465	int	(*dlclose)(void *);
466	void	*(*dlsym)(void *, char *);
467	char	*(*dlerror)(void);
468};
469.Ed
470.Pp
471The
472.Fa crt_ldso
473structure defines the interface between the start-up code in crt0 and
474.Nm ld.so .
475.Bd -literal -offset indent
476struct crt_ldso {
477	int		crt_ba;
478	int		crt_dzfd;
479	int		crt_ldfd;
480	struct _dynamic	*crt_dp;
481	char		**crt_ep;
482	caddr_t		crt_bp;
483	char		*crt_prog;
484	char		*crt_ldso;
485	struct ld_entry	*crt_ldentry;
486};
487#define CRT_VERSION_SUN		1
488#define CRT_VERSION_BSD_2	2
489#define CRT_VERSION_BSD_3	3
490#define	CRT_VERSION_BSD_4	4
491.Ed
492.Bl -tag -width crt_dzfd
493.It Fa crt_ba
494The virtual address at which
495.Nm ld.so
496was loaded by crt0.
497.It Fa crt_dzfd
498On SunOS systems, this field contains an open file descriptor to
499.Dq Pa /dev/zero
500used to get demand paged zeroed pages.
501On
502.Fx
503systems it contains -1.
504.It Fa crt_ldfd
505Contains an open file descriptor that was used by crt0 to load
506.Nm ld.so .
507.It Fa crt_dp
508A pointer to main's
509.Fa _dynamic
510structure.
511.It Fa crt_ep
512A pointer to the environment strings.
513.It Fa crt_bp
514The address at which a breakpoint will be placed by the run-time linker
515if the main program is run by a debugger.
516See
517.Fa so_debug
518.It Fa crt_prog
519The name of the main program as determined by crt0 (CRT_VERSION_BSD3 only).
520.It Fa crt_ldso
521The path of the run-time linker as mapped by crt0 (CRT_VERSION_BSD4 only).
522.El
523.Pp
524The
525.Fa hints_header
526and
527.Fa hints_bucket
528structures define the layout of the library hints, normally found in
529.Dq Pa /var/run/ld.so.hints ,
530which is used by
531.Nm ld.so
532to quickly locate the shared object images in the
533file system.
534The organization of the hints file is not unlike that of an
535.Dq a.out
536object file, in that it contains a header determining the offset and size
537of a table of fixed sized hash buckets and a common string pool.
538.Bd -literal -offset indent
539struct hints_header {
540	long		hh_magic;
541#define HH_MAGIC	011421044151
542	long		hh_version;
543#define LD_HINTS_VERSION_1	1
544	long		hh_hashtab;
545	long		hh_nbucket;
546	long		hh_strtab;
547	long		hh_strtab_sz;
548	long		hh_ehints;
549};
550.Ed
551.Bl -tag -width hh_strtab_sz
552.It Fa hh_magic
553Hints file magic number.
554.It Fa hh_version
555Interface version number.
556.It Fa hh_hashtab
557Offset of hash table.
558.It Fa hh_strtab
559Offset of string table.
560.It Fa hh_strtab_sz
561Size of strings.
562.It Fa hh_ehints
563Maximum usable offset in hints file.
564.El
565.Pp
566.Bd -literal -offset indent
567/*
568 * Hash table element in hints file.
569 */
570struct hints_bucket {
571	int		hi_namex;
572	int		hi_pathx;
573	int		hi_dewey[MAXDEWEY];
574	int		hi_ndewey;
575#define hi_major hi_dewey[0]
576#define hi_minor hi_dewey[1]
577	int		hi_next;
578};
579.Ed
580.Bl -tag -width hi_ndewey
581.It Fa hi_namex
582Index of the string identifying the library.
583.It Fa hi_pathx
584Index of the string representing the full path name of the library.
585.It Fa hi_dewey
586The version numbers of the shared library.
587.It Fa hi_ndewey
588The number of valid entries in
589.Fa hi_dewey .
590.It Fa hi_next
591Next bucket in case of hashing collisions.
592.El
593.Sh CAVEATS
594Only the (GNU) C compiler currently supports the creation of shared libraries.
595Other programming languages cannot be used.
596