The /var/run/ld.so.hints file is mapped into memory. It is never replaced,
so the mapping can be immutable.
ok kettenis

Prep .c files for removing the #includes from */archdep.h
* replace #include "archdep.h" with #includes of what is used, pulling in
"syscall.h", "util.h", and "archdep.h" as needed
* delete #include <sys/syscall.h> from syscall.h
* only pull in <sys/stat.h> to the three files that use _dl_fstat(),
forward declare struct stat in syscall.h for the others
* NBBY is for <sys/select.h> macros; just use '8' in dl_printf.c
* <machine/vmparam.h> is only needed on i386; conditionalize it
* stop using __LDPGSZ: use _MAX_PAGE_SHIFT (already used by malloc.c)
where necessary
* delete other bogus #includes, order legit per style: <sys/*> then
<*/*>, then <*>, then "*"

dir.c improvement from jsg@
ok and testing assistance deraadt@

ld.so's syscall _dl_* API/ABI doesn't reurn errno, but -errno.
Reviewing behaviour artifacts with pledge, I discover a close(-ENOSYS)..
ok guenther

On fatal errors, kill ourselves with thrkill(0,9,NULL) instead of
simply exiting, via helper functions _dl_die(), _dl_diedie(), and
_dl_oom().

prompted by a complaint from jsing@
ok jsing@ deraadt@

remove tests for LD_HINTS_VERSION_1; ok kettenis

Nuke some obvious #include duplications.

ok espie@ deraadt@ millert@ tedu@

check all memory allocations; ok miod@ guenther@

Use O_CLOEXEC to make sure fork+exec in a threaded process can't
see the fds used by dlopen()

ok otto@ miod@

enable -Wall, explicitly call _dl_strsep, and zap some redundant includes
ok miod@

Move to a non-zeroing _dl_malloc, a _dl_calloc and _dl_reallocarry and
fix _dl_strdup to return NULL instead of crash; ok deraadt@

stop using sys/syslimits.h directly, and use limits.h
prodded by millert

- Parse colon separated paths into NULL terminated arrays of string
pointers to prepare for adding rpath ORIGIN support.
okay matthew@ millert@

Do not divide by zero in _dl_findhint() if there is a valid hints file but
no buckets in it (could happen if the administrator reruns ldconfig with
bogus paths). ok matthew@ kurt@ drahn

Fix loaded object sod matching: when we load libfoo.so.X.Y into
memory, we should be able to match other requests for libfoo.so.X.Z
against that same object.

ok kurt, kettenis

Fix mmap() error checking to be correct 64-bit addresses. Consistently
use _dl_mmap_error() to check for mmap() errors. Adjust datatypes of
some local vars for 64-bit safety.

okay millert@ drahn@

Fix some problems related to LD_LIBRARY_PATH parsing where it would not
correctly deal with current directory searches specified by "::", ":foo" or
"foo:"

fix includes scoping; drahn@ ok

Better fix to the library name parsing. now correctly handles
libpython2.2.so.0.0 vs libpython2.1.so.0.0. Clean up a potential
memory leak if the filename is not a valid library.
ok brad@

Work around a problem with shared libraries named as libX-A.B.so.X.Y.
Without this patch, the standard major:minor version matching failed.
ok brad@

various proto, ansi, and knf repair. tested on all architectures that
use it. (build may require make cleandir because of .depend balony)

Previously, ld.so would use a single page for mapping the hints file
and then, if that was not big enough, map the rest of it contiguously.
However, there is no guarantee that the address space after the
first mmap has not already been used by something else.

Instead, just fstat the file and map up to st_size. Also reorganized
the code a tad and added some sanity checks for st_size.

Problem tracked down by dhartmei@ and mickey@; drahn@ OK.

branches: 1.16.2;
knf & ansi; drahn ok

minor formatting

Change ld.so search order/method to match the a.out ld.so.

run destructors on dlclose()

Move more symbols into _dl_ private space, so that the proper (libc)
version of the function will be used.

Add readdir() functionality to perform the proper library searching.

Support DL_PRELOAD

Do not relocate symbols if ld.so is being traced (and will exit).

Misc lint cleanup.

ok art@

KNF

use _dl_strncmp() not strncmp(), _dl version is already present and used.

_dl_strcmp() and _dl_strchr() exist, do not pull these routines
from libsa source.

more KNF

various KNF

cleanup of MD/MI ld.so (elf), most changes by art@ cleanup by me.
tested on alpha sparc64 powerpc.

Upon futher contemplation, the notion of unmaphints is just wrong.
First the maphints code is not cleaning up after itself, by leaving
a file descriptor open and trying to clean up after the dynamic linker
is done loading by removing data needed during dlopen() is nonsense.
the _dl_unmaphints() function and it's references have been removed.
_dl_maphints() now cleans up it's file descriptor before returning.
static data used later during dlopen() is left intact.

The unmaphints routine should not destroy the hint mapping.
Otherwise shared modules which depend on libraries cannot find
them because the list of trusted directories has been lost.
However we need to be careful that the unmapping of the hints frees
the file descriptor, but will not free it multiple times.

Be consistent. If we have inveted our of strlen, use it everywhere.

When opening the hints file, check for error return < 0, not -1.
_dl_open is not necessarily like the regular open(2).
XXX - maybe we should check with _dl_check_error?

MAP_COPY -> MAP_PRIVATE

-1 -> MAP_FAILED.

branches: 1.1.1;
Initial revision

Prep .c files for removing the #includes from */archdep.h
* replace #include "archdep.h" with #includes of what is used, pulling in
"syscall.h", "util.h", and "archdep.h" as needed
* delete #include <sys/syscall.h> from syscall.h
* only pull in <sys/stat.h> to the three files that use _dl_fstat(),
forward declare struct stat in syscall.h for the others
* NBBY is for <sys/select.h> macros; just use '8' in dl_printf.c
* <machine/vmparam.h> is only needed on i386; conditionalize it
* stop using __LDPGSZ: use _MAX_PAGE_SHIFT (already used by malloc.c)
where necessary
* delete other bogus #includes, order legit per style: <sys/*> then
<*/*>, then <*>, then "*"

dir.c improvement from jsg@
ok and testing assistance deraadt@

ld.so's syscall _dl_* API/ABI doesn't reurn errno, but -errno.
Reviewing behaviour artifacts with pledge, I discover a close(-ENOSYS)..
ok guenther

On fatal errors, kill ourselves with thrkill(0,9,NULL) instead of
simply exiting, via helper functions _dl_die(), _dl_diedie(), and
_dl_oom().

prompted by a complaint from jsing@
ok jsing@ deraadt@

remove tests for LD_HINTS_VERSION_1; ok kettenis

Nuke some obvious #include duplications.

ok espie@ deraadt@ millert@ tedu@

check all memory allocations; ok miod@ guenther@

Use O_CLOEXEC to make sure fork+exec in a threaded process can't
see the fds used by dlopen()

ok otto@ miod@

enable -Wall, explicitly call _dl_strsep, and zap some redundant includes
ok miod@

Move to a non-zeroing _dl_malloc, a _dl_calloc and _dl_reallocarry and
fix _dl_strdup to return NULL instead of crash; ok deraadt@

stop using sys/syslimits.h directly, and use limits.h
prodded by millert

- Parse colon separated paths into NULL terminated arrays of string
pointers to prepare for adding rpath ORIGIN support.
okay matthew@ millert@

Do not divide by zero in _dl_findhint() if there is a valid hints file but
no buckets in it (could happen if the administrator reruns ldconfig with
bogus paths). ok matthew@ kurt@ drahn

Fix loaded object sod matching: when we load libfoo.so.X.Y into
memory, we should be able to match other requests for libfoo.so.X.Z
against that same object.

ok kurt, kettenis

Fix mmap() error checking to be correct 64-bit addresses. Consistently
use _dl_mmap_error() to check for mmap() errors. Adjust datatypes of
some local vars for 64-bit safety.

okay millert@ drahn@

Fix some problems related to LD_LIBRARY_PATH parsing where it would not
correctly deal with current directory searches specified by "::", ":foo" or
"foo:"

fix includes scoping; drahn@ ok

Better fix to the library name parsing. now correctly handles
libpython2.2.so.0.0 vs libpython2.1.so.0.0. Clean up a potential
memory leak if the filename is not a valid library.
ok brad@

Work around a problem with shared libraries named as libX-A.B.so.X.Y.
Without this patch, the standard major:minor version matching failed.
ok brad@

various proto, ansi, and knf repair. tested on all architectures that
use it. (build may require make cleandir because of .depend balony)

Previously, ld.so would use a single page for mapping the hints file
and then, if that was not big enough, map the rest of it contiguously.
However, there is no guarantee that the address space after the
first mmap has not already been used by something else.

Instead, just fstat the file and map up to st_size. Also reorganized
the code a tad and added some sanity checks for st_size.

Problem tracked down by dhartmei@ and mickey@; drahn@ OK.

branches: 1.16.2;
knf & ansi; drahn ok

minor formatting

Change ld.so search order/method to match the a.out ld.so.

run destructors on dlclose()

Move more symbols into _dl_ private space, so that the proper (libc)
version of the function will be used.

Add readdir() functionality to perform the proper library searching.

Support DL_PRELOAD

Do not relocate symbols if ld.so is being traced (and will exit).

Misc lint cleanup.

ok art@

KNF

use _dl_strncmp() not strncmp(), _dl version is already present and used.

_dl_strcmp() and _dl_strchr() exist, do not pull these routines
from libsa source.

more KNF

various KNF

cleanup of MD/MI ld.so (elf), most changes by art@ cleanup by me.
tested on alpha sparc64 powerpc.

Upon futher contemplation, the notion of unmaphints is just wrong.
First the maphints code is not cleaning up after itself, by leaving
a file descriptor open and trying to clean up after the dynamic linker
is done loading by removing data needed during dlopen() is nonsense.
the _dl_unmaphints() function and it's references have been removed.
_dl_maphints() now cleans up it's file descriptor before returning.
static data used later during dlopen() is left intact.

The unmaphints routine should not destroy the hint mapping.
Otherwise shared modules which depend on libraries cannot find
them because the list of trusted directories has been lost.
However we need to be careful that the unmapping of the hints frees
the file descriptor, but will not free it multiple times.

Be consistent. If we have inveted our of strlen, use it everywhere.

When opening the hints file, check for error return < 0, not -1.
_dl_open is not necessarily like the regular open(2).
XXX - maybe we should check with _dl_check_error?

MAP_COPY -> MAP_PRIVATE

-1 -> MAP_FAILED.

branches: 1.1.1;
Initial revision

ld.so's syscall _dl_* API/ABI doesn't reurn errno, but -errno.
Reviewing behaviour artifacts with pledge, I discover a close(-ENOSYS)..
ok guenther

On fatal errors, kill ourselves with thrkill(0,9,NULL) instead of
simply exiting, via helper functions _dl_die(), _dl_diedie(), and
_dl_oom().

prompted by a complaint from jsing@
ok jsing@ deraadt@

remove tests for LD_HINTS_VERSION_1; ok kettenis

Nuke some obvious #include duplications.

ok espie@ deraadt@ millert@ tedu@

check all memory allocations; ok miod@ guenther@

Use O_CLOEXEC to make sure fork+exec in a threaded process can't
see the fds used by dlopen()

ok otto@ miod@

enable -Wall, explicitly call _dl_strsep, and zap some redundant includes
ok miod@

Move to a non-zeroing _dl_malloc, a _dl_calloc and _dl_reallocarry and
fix _dl_strdup to return NULL instead of crash; ok deraadt@

stop using sys/syslimits.h directly, and use limits.h
prodded by millert

- Parse colon separated paths into NULL terminated arrays of string
pointers to prepare for adding rpath ORIGIN support.
okay matthew@ millert@

Do not divide by zero in _dl_findhint() if there is a valid hints file but
no buckets in it (could happen if the administrator reruns ldconfig with
bogus paths). ok matthew@ kurt@ drahn

Fix loaded object sod matching: when we load libfoo.so.X.Y into
memory, we should be able to match other requests for libfoo.so.X.Z
against that same object.

ok kurt, kettenis

Fix mmap() error checking to be correct 64-bit addresses. Consistently
use _dl_mmap_error() to check for mmap() errors. Adjust datatypes of
some local vars for 64-bit safety.

okay millert@ drahn@

Fix some problems related to LD_LIBRARY_PATH parsing where it would not
correctly deal with current directory searches specified by "::", ":foo" or
"foo:"

fix includes scoping; drahn@ ok

Better fix to the library name parsing. now correctly handles
libpython2.2.so.0.0 vs libpython2.1.so.0.0. Clean up a potential
memory leak if the filename is not a valid library.
ok brad@

Work around a problem with shared libraries named as libX-A.B.so.X.Y.
Without this patch, the standard major:minor version matching failed.
ok brad@

various proto, ansi, and knf repair. tested on all architectures that
use it. (build may require make cleandir because of .depend balony)

Previously, ld.so would use a single page for mapping the hints file
and then, if that was not big enough, map the rest of it contiguously.
However, there is no guarantee that the address space after the
first mmap has not already been used by something else.

Instead, just fstat the file and map up to st_size. Also reorganized
the code a tad and added some sanity checks for st_size.

Problem tracked down by dhartmei@ and mickey@; drahn@ OK.

branches: 1.16.2;
knf & ansi; drahn ok

minor formatting

Change ld.so search order/method to match the a.out ld.so.

run destructors on dlclose()

Move more symbols into _dl_ private space, so that the proper (libc)
version of the function will be used.

Add readdir() functionality to perform the proper library searching.

Support DL_PRELOAD

Do not relocate symbols if ld.so is being traced (and will exit).

Misc lint cleanup.

ok art@

KNF

use _dl_strncmp() not strncmp(), _dl version is already present and used.

_dl_strcmp() and _dl_strchr() exist, do not pull these routines
from libsa source.

more KNF

various KNF

cleanup of MD/MI ld.so (elf), most changes by art@ cleanup by me.
tested on alpha sparc64 powerpc.

Upon futher contemplation, the notion of unmaphints is just wrong.
First the maphints code is not cleaning up after itself, by leaving
a file descriptor open and trying to clean up after the dynamic linker
is done loading by removing data needed during dlopen() is nonsense.
the _dl_unmaphints() function and it's references have been removed.
_dl_maphints() now cleans up it's file descriptor before returning.
static data used later during dlopen() is left intact.

The unmaphints routine should not destroy the hint mapping.
Otherwise shared modules which depend on libraries cannot find
them because the list of trusted directories has been lost.
However we need to be careful that the unmapping of the hints frees
the file descriptor, but will not free it multiple times.

Be consistent. If we have inveted our of strlen, use it everywhere.

When opening the hints file, check for error return < 0, not -1.
_dl_open is not necessarily like the regular open(2).
XXX - maybe we should check with _dl_check_error?

MAP_COPY -> MAP_PRIVATE

-1 -> MAP_FAILED.

branches: 1.1.1;
Initial revision