freebsd-dev/sys/kern/link_aout.c

552 lines
14 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1997-2000 Doug Rabson
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
1999-08-28 01:08:13 +00:00
* $FreeBSD$
*/
#ifdef __i386__
1998-08-16 04:19:03 +00:00
#define FREEBSD_AOUT 1
#include <sys/param.h>
#include <sys/kernel.h>
#include <sys/systm.h>
#include <sys/malloc.h>
#include <sys/proc.h>
#include <sys/namei.h>
#include <sys/fcntl.h>
#include <sys/vnode.h>
#include <sys/linker.h>
#include "linker_if.h"
#ifndef __ELF__
#include <vm/vm.h>
#include <vm/pmap.h>
#include <machine/vmparam.h>
#endif
#include <a.out.h>
#include <link.h>
typedef struct aout_file {
struct linker_file lf; /* Common fields */
int preloaded; /* Was this pre-loader */
char* address; /* Load address */
struct _dynamic* dynamic; /* Symbol table etc. */
} *aout_file_t;
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
static int link_aout_link_preload(linker_class_t lc,
const char* modname, linker_file_t*);
static int link_aout_link_preload_finish(linker_file_t);
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
static int link_aout_load_file(linker_class_t lc, const char*, linker_file_t*);
static int link_aout_lookup_symbol(linker_file_t, const char*,
c_linker_sym_t*);
static int link_aout_symbol_values(linker_file_t file, c_linker_sym_t sym,
linker_symval_t* symval);
static int link_aout_search_symbol(linker_file_t lf, caddr_t value,
c_linker_sym_t* sym, long* diffp);
static void link_aout_unload_file(linker_file_t);
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
static void link_aout_unload_preload(linker_file_t);
static kobj_method_t link_aout_methods[] = {
KOBJMETHOD(linker_lookup_symbol, link_aout_lookup_symbol),
KOBJMETHOD(linker_symbol_values, link_aout_symbol_values),
KOBJMETHOD(linker_search_symbol, link_aout_search_symbol),
KOBJMETHOD(linker_unload, link_aout_unload_file),
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
KOBJMETHOD(linker_load_file, link_aout_load_file),
KOBJMETHOD(linker_link_preload, link_aout_link_preload),
KOBJMETHOD(linker_link_preload_finish, link_aout_link_preload_finish),
{ 0, 0 }
};
static struct linker_class link_aout_class = {
"a.out", link_aout_methods, sizeof(struct aout_file)
};
static int relocate_file(aout_file_t af);
/*
* The kernel symbol table starts here.
*/
extern struct _dynamic _DYNAMIC;
static void
link_aout_init(void* arg)
{
#ifndef __ELF__
struct _dynamic* dp = &_DYNAMIC;
#endif
linker_add_class(&link_aout_class);
#ifndef __ELF__
if (dp) {
aout_file_t af;
linker_kernel_file =
linker_make_file(kernelname, af, &link_aout_class);
if (linker_kernel_file == NULL)
panic("link_aout_init: Can't create linker structures for kernel");
af = (aout_file_t) linker_kernel_file;
af->address = 0;
af->dynamic = dp;
linker_kernel_file->address = (caddr_t) KERNBASE;
linker_kernel_file->size = -(long)linker_kernel_file->address;
}
#endif
}
SYSINIT(link_aout, SI_SUB_KLD, SI_ORDER_THIRD, link_aout_init, 0);
static int
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
link_aout_link_preload(linker_class_t lc,
const char* filename, linker_file_t* result)
{
caddr_t modptr, baseptr;
char *type;
struct exec *ehdr;
aout_file_t af;
linker_file_t lf;
/* Look to see if we have the module preloaded. */
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
modptr = preload_search_by_name(filename);
if (modptr == NULL)
return ENOENT;
if (((type = (char *)preload_search_info(modptr, MODINFO_TYPE)) == NULL) ||
strcmp(type, "a.out module") ||
((baseptr = preload_search_info(modptr, MODINFO_ADDR)) == NULL) ||
((ehdr = (struct exec *)preload_search_info(modptr, MODINFO_METADATA | MODINFOMD_AOUTEXEC)) == NULL))
return(0); /* we can't handle this */
/* Register with kld */
lf = linker_make_file(filename, &link_aout_class);
if (lf == NULL) {
return(ENOMEM);
}
af = (aout_file_t) lf;
/* Looks like we can handle this one */
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
filename = preload_search_info(modptr, MODINFO_NAME);
af->preloaded = 1;
af->address = baseptr;
/* Assume _DYNAMIC is the first data item. */
af->dynamic = (struct _dynamic*)(af->address + ehdr->a_text);
if (af->dynamic->d_version != LD_VERSION_BSD) {
linker_file_unload(lf);
return(0); /* we can't handle this */
}
af->dynamic->d_un.d_sdt = (struct section_dispatch_table *)
((char *)af->dynamic->d_un.d_sdt + (vm_offset_t)af->address);
lf->address = af->address;
lf->size = ehdr->a_text + ehdr->a_data + ehdr->a_bss;
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
*result = lf;
return(0);
}
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
static int
link_aout_link_preload_finish(linker_file_t lf)
{
aout_file_t af;
int error;
af = (aout_file_t) lf;
error = relocate_file(af);
if (error) {
linker_file_unload(lf);
return(error);
}
return(0);
}
static int
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
link_aout_load_file(linker_class_t lc, const char* filename, linker_file_t* result)
{
struct nameidata nd;
struct proc* p = curproc; /* XXX */
int error = 0;
int resid, flags;
struct exec header;
aout_file_t af;
linker_file_t lf = 0;
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
NDINIT(&nd, LOOKUP, FOLLOW, UIO_SYSSPACE, filename, p);
flags = FREAD;
error = vn_open(&nd, &flags, 0);
if (error)
return error;
NDFREE(&nd, NDF_ONLY_PNBUF);
/*
* Read the a.out header from the file.
*/
error = vn_rdwr(UIO_READ, nd.ni_vp, (void*) &header, sizeof header, 0,
UIO_SYSSPACE, IO_NODELOCKED, p->p_ucred, &resid, p);
if (error)
goto out;
if (N_BADMAG(header) || !(N_GETFLAG(header) & EX_DYNAMIC))
goto out;
/*
* We have an a.out file, so make some space to read it in.
*/
lf = linker_make_file(filename, &link_aout_class);
if (lf == NULL) {
error = ENOMEM;
goto out;
}
af = (aout_file_t) lf;
af->address = malloc(header.a_text + header.a_data + header.a_bss,
M_LINKER, M_WAITOK);
/*
* Read the text and data sections and zero the bss.
*/
error = vn_rdwr(UIO_READ, nd.ni_vp, (void*) af->address,
header.a_text + header.a_data, 0,
UIO_SYSSPACE, IO_NODELOCKED, p->p_ucred, &resid, p);
if (error)
goto out;
bzero(af->address + header.a_text + header.a_data, header.a_bss);
/*
* Assume _DYNAMIC is the first data item.
*/
af->dynamic = (struct _dynamic*) (af->address + header.a_text);
if (af->dynamic->d_version != LD_VERSION_BSD) {
error = ENOEXEC;
goto out;
}
af->dynamic->d_un.d_sdt = (struct section_dispatch_table *)
((char *)af->dynamic->d_un.d_sdt + (vm_offset_t)af->address);
lf->address = af->address;
lf->size = header.a_text + header.a_data + header.a_bss;
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
error = linker_load_dependancies(lf);
if (error)
goto out;
error = relocate_file(af);
if (error)
goto out;
*result = lf;
out:
if (error && lf)
linker_file_unload(lf);
VOP_UNLOCK(nd.ni_vp, 0, p);
vn_close(nd.ni_vp, FREAD, p->p_ucred, p);
return error;
}
static void
link_aout_unload_file(linker_file_t file)
{
aout_file_t af = (aout_file_t) file;
if (af->preloaded) {
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
link_aout_unload_preload(file);
return;
}
if (af->address)
free(af->address, M_LINKER);
}
static void
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
link_aout_unload_preload(linker_file_t file)
{
if (file->filename)
preload_delete_name(file->filename);
}
/*
* XXX i386 dependant.
*/
static long
read_relocation(struct relocation_info* r, char* addr)
{
int length = r->r_length;
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
if (length == 0)
return *(u_char*) addr;
else if (length == 1)
return *(u_short*) addr;
else if (length == 2)
return *(u_int*) addr;
else
printf("link_aout: unsupported relocation size %d\n", r->r_length);
return 0;
}
static void
write_relocation(struct relocation_info* r, char* addr, long value)
{
int length = r->r_length;
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
if (length == 0)
*(u_char*) addr = value;
else if (length == 1)
*(u_short*) addr = value;
else if (length == 2)
*(u_int*) addr = value;
else
printf("link_aout: unsupported relocation size %d\n", r->r_length);
}
First round implementation of a fine grain enhanced module to module version dependency system. This isn't quite finished, but it is at a useful stage to do a functional checkpoint. Highlights: - version and dependency metadata is gathered via linker sets, so things are handled the same for static kernels and code built to live in a kld. - The dependencies are at module level (versus at file level). - Dependencies determine kld symbol search order - this means that you cannot link against symbols in another file unless you depend on it. This is so that you cannot accidently unload the target out from underneath the ones referencing it. - It is flexible enough that we can put tags in #include files and macros so that we can get decent hooks for enforcing recompiles on incompatable ABI changes. eg: if we change struct proc, we could force a recompile for all kld's that reference the proc struct. - Tangled dependency references at boot time are sorted. Files are relocated once all their dependencies are already relocated. Caveats: - Loader support is incomplete, but has been worked on seperately. - Actual enforcement of the version number tags is not active yet - just the module dependencies are live. The actual structure of versioning hasn't been agreed on yet. (eg: major.minor, or whatever) - There is some backwards compatability for old modules without metadata but I'm not sure how good it is. This is based on work originally done by Boris Popov (bp@freebsd.org), but I'm not sure he'd recognize much of it now. Don't blame him. :-) Also, ideas have been borrowed from Mike Smith.
2000-04-29 13:19:31 +00:00
#define AOUT_RELOC(af, type, off) (type*) ((af)->address + (off))
static int
relocate_file(aout_file_t af)
{
struct relocation_info* rel;
struct relocation_info* erel;
struct relocation_info* r;
struct nzlist* symbolbase;
char* stringbase;
struct nzlist* np;
char* sym;
long relocation;
rel = AOUT_RELOC(af, struct relocation_info, LD_REL(af->dynamic));
erel = AOUT_RELOC(af, struct relocation_info,
LD_REL(af->dynamic) + LD_RELSZ(af->dynamic));
symbolbase = AOUT_RELOC(af, struct nzlist, LD_SYMBOL(af->dynamic));
stringbase = AOUT_RELOC(af, char, LD_STRINGS(af->dynamic));
for (r = rel; r < erel; r++) {
char* addr;
if (r->r_address == 0)
break;
addr = AOUT_RELOC(af, char, r->r_address);
if (r->r_extern) {
np = &symbolbase[r->r_symbolnum];
sym = &stringbase[np->nz_strx];
if (sym[0] != '_') {
printf("link_aout: bad symbol name %s\n", sym);
relocation = 0;
} else
relocation = (intptr_t)
linker_file_lookup_symbol(&af->lf, sym + 1,
np->nz_type != (N_SETV+N_EXT));
if (!relocation) {
printf("link_aout: symbol %s not found\n", sym);
return ENOENT;
}
relocation += read_relocation(r, addr);
if (r->r_jmptable) {
printf("link_aout: can't cope with jump table relocations\n");
continue;
}
if (r->r_pcrel)
relocation -= (intptr_t) af->address;
if (r->r_copy) {
printf("link_aout: can't cope with copy relocations\n");
continue;
}
write_relocation(r, addr, relocation);
} else {
write_relocation(r, addr,
(intptr_t)(read_relocation(r, addr) + af->address));
}
}
return 0;
}
static long
symbol_hash_value(aout_file_t af, const char* name)
{
long hashval;
const char* p;
hashval = '_'; /* fake a starting '_' for C symbols */
for (p = name; *p; p++)
hashval = (hashval << 1) + *p;
return (hashval & 0x7fffffff) % LD_BUCKETS(af->dynamic);
}
int
link_aout_lookup_symbol(linker_file_t file, const char* name,
c_linker_sym_t* sym)
{
aout_file_t af = (aout_file_t) file;
long hashval;
struct rrs_hash* hashbase;
struct nzlist* symbolbase;
char* stringbase;
struct rrs_hash* hp;
struct nzlist* np;
char* cp;
if (LD_BUCKETS(af->dynamic) == 0)
1999-12-21 11:14:12 +00:00
return 0;
hashbase = AOUT_RELOC(af, struct rrs_hash, LD_HASH(af->dynamic));
symbolbase = AOUT_RELOC(af, struct nzlist, LD_SYMBOL(af->dynamic));
stringbase = AOUT_RELOC(af, char, LD_STRINGS(af->dynamic));
restart:
hashval = symbol_hash_value(af, name);
hp = &hashbase[hashval];
if (hp->rh_symbolnum == -1)
return ENOENT;
while (hp) {
np = (struct nzlist *) &symbolbase[hp->rh_symbolnum];
cp = stringbase + np->nz_strx;
/*
* Note: we fake the leading '_' for C symbols.
*/
if (cp[0] == '_' && !strcmp(cp + 1, name))
break;
if (hp->rh_next == 0)
hp = NULL;
else
hp = &hashbase[hp->rh_next];
}
if (hp == NULL)
/*
* Not found.
*/
return ENOENT;
/*
* Check for an aliased symbol, whatever that is.
*/
if (np->nz_type == N_INDR+N_EXT) {
name = stringbase + (++np)->nz_strx + 1; /* +1 for '_' */
goto restart;
}
/*
* Check this is an actual definition of the symbol.
*/
if (np->nz_value == 0)
return ENOENT;
if (np->nz_type == N_UNDF+N_EXT && np->nz_value != 0) {
if (np->nz_other == AUX_FUNC)
/* weak function */
return ENOENT;
}
*sym = (linker_sym_t) np;
return 0;
}
static int
link_aout_symbol_values(linker_file_t file, c_linker_sym_t sym,
linker_symval_t* symval)
{
aout_file_t af = (aout_file_t) file;
const struct nzlist* np = (const struct nzlist*) sym;
char* stringbase;
long numsym = LD_STABSZ(af->dynamic) / sizeof(struct nzlist);
struct nzlist *symbase;
/* Is it one of ours? It could be another module... */
symbase = AOUT_RELOC(af, struct nzlist, LD_SYMBOL(af->dynamic));
if (np < symbase)
return ENOENT;
if ((np - symbase) > numsym)
return ENOENT;
stringbase = AOUT_RELOC(af, char, LD_STRINGS(af->dynamic));
symval->name = stringbase + np->nz_strx + 1; /* +1 for '_' */
if (np->nz_type == N_UNDF+N_EXT && np->nz_value != 0) {
symval->value = 0;
symval->size = np->nz_value;
} else {
symval->value = AOUT_RELOC(af, char, np->nz_value);
symval->size = np->nz_size;
}
return 0;
}
static int
link_aout_search_symbol(linker_file_t lf, caddr_t value,
c_linker_sym_t* sym, long* diffp)
{
aout_file_t af = (aout_file_t) lf;
u_long off = (uintptr_t) (void *) value;
u_long diff = off;
u_long sp_nz_value;
struct nzlist* sp;
struct nzlist* ep;
struct nzlist* best = 0;
for (sp = AOUT_RELOC(af, struct nzlist, LD_SYMBOL(af->dynamic)),
ep = (struct nzlist *) ((caddr_t) sp + LD_STABSZ(af->dynamic));
sp < ep; sp++) {
if (sp->nz_name == 0)
continue;
sp_nz_value = sp->nz_value + (uintptr_t) (void *) af->address;
if (off >= sp_nz_value) {
if (off - sp_nz_value < diff) {
diff = off - sp_nz_value;
best = sp;
if (diff == 0)
break;
} else if (off - sp_nz_value == diff) {
best = sp;
}
}
}
if (best == 0)
*diffp = off;
else
*diffp = diff;
*sym = (linker_sym_t) best;
return 0;
}
#endif /* !__alpha__ */