This reverts commit 6985865bc3.
Reason for revert:
The commit changes the order of ELF destructor calls too much relative
to what applications expect or can handle. In particular, during
process exit and _dl_fini, after the revert commit, we no longer call
the destructors of the main program first; that only happens after
some dlopen'ed objects have been destructed. This robs applications
of an opportunity to influence destructor order by calling dlclose
explicitly from the main program's ELF destructors. A couple of
different approaches involving reverse constructor order were tried,
and none of them worked really well. It seems we need to keep the
dependency sorting in _dl_fini.
There is also an ambiguity regarding nested dlopen calls from ELF
constructors: Should those destructors run before or after the object
that called dlopen? Commit 6985865bc3 used reverse order
of the start of ELF constructor calls for destructors, but arguably
using completion of constructors is more correct. However, that alone
is not sufficient to address application compatibility issues (it
does not change _dl_fini ordering at all).
The current implementation of dlclose (and process exit) re-sorts the
link maps before calling ELF destructors. Destructor order is not the
reverse of the constructor order as a result: The second sort takes
relocation dependencies into account, and other differences can result
from ambiguous inputs, such as cycles. (The force_first handling in
_dl_sort_maps is not effective for dlclose.) After the changes in
this commit, there is still a required difference due to
dlopen/dlclose ordering by the application, but the previous
discrepancies went beyond that.
A new global (namespace-spanning) list of link maps,
_dl_init_called_list, is updated right before ELF constructors are
called from _dl_init.
In dl_close_worker, the maps variable, an on-stack variable length
array, is eliminated. (VLAs are problematic, and dlclose should not
call malloc because it cannot readily deal with malloc failure.)
Marking still-used objects uses the namespace list directly, with
next and next_idx replacing the done_index variable.
After marking, _dl_init_called_list is used to call the destructors
of now-unused maps in reverse destructor order. These destructors
can call dlopen. Previously, new objects do not have l_map_used set.
This had to change: There is no copy of the link map list anymore,
so processing would cover newly opened (and unmarked) mappings,
unloading them. Now, _dl_init (indirectly) sets l_map_used, too.
(dlclose is handled by the existing reentrancy guard.)
After _dl_init_called_list traversal, two more loops follow. The
processing order changes to the original link map order in the
namespace. Previously, dependency order was used. The difference
should not matter because relocation dependencies could already
reorder link maps in the old code.
The changes to _dl_fini remove the sorting step and replace it with
a traversal of _dl_init_called_list. The l_direct_opencount
decrement outside the loader lock is removed because it appears
incorrect: the counter manipulation could race with other dynamic
loader operations.
tst-audit23 needs adjustments to the changes in LA_ACT_DELETE
notifications. The new approach for checking la_activity should
make it clearer that la_activty calls come in pairs around namespace
updates.
The dependency sorting test cases need updates because the destructor
order is always the opposite order of constructor order, even with
relocation dependencies or cycles present.
There is a future cleanup opportunity to remove the now-constant
force_first and for_fini arguments from the _dl_sort_maps function.
Fixes commit 1df71d32fe ("elf: Implement
force_first handling in _dl_sort_maps_dfs (bug 28937)").
Reviewed-by: DJ Delorie <dj@redhat.com>
This consolidates the destructor invocations from _dl_fini and
dlclose. Remove the micro-optimization that avoids
calling _dl_call_fini if they are no destructors (as dlclose is quite
expensive anyway). The debug log message is now printed
unconditionally.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
la_activity is not called during application exit, even though
la_objclose is.
Checked on x86_64-linux-gnu, i686-linux-gnu, and aarch64-linux-gnu.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
I used these shell commands:
../glibc/scripts/update-copyrights $PWD/../gnulib/build-aux/update-copyright
(cd ../glibc && git commit -am"[this commit message]")
and then ignored the output, which consisted lines saying "FOO: warning:
copyright statement not found" for each of 7061 files FOO.
I then removed trailing white space from math/tgmath.h,
support/tst-support-open-dev-null-range.c, and
sysdeps/x86_64/multiarch/strlen-vec.S, to work around the following
obscure pre-commit check failure diagnostics from Savannah. I don't
know why I run into these diagnostics whereas others evidently do not.
remote: *** 912-#endif
remote: *** 913:
remote: *** 914-
remote: *** error: lines with trailing whitespace found
...
remote: *** error: sysdeps/unix/sysv/linux/statx_cp.c: trailing lines
It consolidates the code required to call la_objclose audit
callback.
Checked on x86_64-linux-gnu, i686-linux-gnu, and aarch64-linux-gnu.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
This second patch contains the actual implementation of a new sorting algorithm
for shared objects in the dynamic loader, which solves the slow behavior that
the current "old" algorithm falls into when the DSO set contains circular
dependencies.
The new algorithm implemented here is simply depth-first search (DFS) to obtain
the Reverse-Post Order (RPO) sequence, a topological sort. A new l_visited:1
bitfield is added to struct link_map to more elegantly facilitate such a search.
The DFS algorithm is applied to the input maps[nmap-1] backwards towards
maps[0]. This has the effect of a more "shallow" recursion depth in general
since the input is in BFS. Also, when combined with the natural order of
processing l_initfini[] at each node, this creates a resulting output sorting
closer to the intuitive "left-to-right" order in most cases.
Another notable implementation adjustment related to this _dl_sort_maps change
is the removing of two char arrays 'used' and 'done' in _dl_close_worker to
represent two per-map attributes. This has been changed to simply use two new
bit-fields l_map_used:1, l_map_done:1 added to struct link_map. This also allows
discarding the clunky 'used' array sorting that _dl_sort_maps had to sometimes
do along the way.
Tunable support for switching between different sorting algorithms at runtime is
also added. A new tunable 'glibc.rtld.dynamic_sort' with current valid values 1
(old algorithm) and 2 (new DFS algorithm) has been added. At time of commit
of this patch, the default setting is 1 (old algorithm).
Signed-off-by: Chung-Lin Tang <cltang@codesourcery.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
I used these shell commands:
../glibc/scripts/update-copyrights $PWD/../gnulib/build-aux/update-copyright
(cd ../glibc && git commit -am"[this commit message]")
and then ignored the output, which consisted lines saying "FOO: warning:
copyright statement not found" for each of 6694 files FOO.
I then removed trailing white space from benchtests/bench-pthread-locks.c
and iconvdata/tst-iconv-big5-hkscs-to-2ucs4.c, to work around this
diagnostic from Savannah:
remote: *** pre-commit check failed ...
remote: *** error: lines with trailing whitespace found
remote: error: hook declined to update refs/heads/master
This supersedes the init_array sysdeps directory. It allows us to
check for ELF_INITFINI in both C and assembler code, and skip DT_INIT
and DT_FINI processing completely on newer architectures.
A new header file is needed because <dl-machine.h> is incompatible
with assembler code. <sysdep.h> is compatible with assembler code,
but it cannot be included in all assembler files because on some
architectures, it redefines register names, and some assembler files
conflict with that.
<elf-initfini.h> is replicated for legacy architectures which need
DT_INIT/DT_FINI support. New architectures follow the generic default
and disable it.
To improve GCC 10 compatibility, it is necessary to remove the l_audit
zero-length array from the end of struct link_map. In preparation of
that, this commit introduces an accessor function for the audit state,
so that it is possible to change the representation of the audit state
without adjusting the code that accesses it.
Tested on x86_64-linux-gnu. Built on i686-gnu.
Change-Id: Id815673c29950fc011ae5301d7cde12624f658df
Combine the four places where link maps are sorted into a single function.
This also moves the logic to skip the first map (representing the main
binary) to the callers.
Assembler code passes the address of _dl_fini to __libc_start_main,
whose function pointer argument lacks the attribute. This means
that calls could use the wrong ABI. Fortunately, for zero-parameter
void-returning functions, internal_function does not change ABI
on i386 (the only architecture which uses internal_function), so
this inconsistency was harmless (which is why it had not been
noticed so far).
In this case, extend_alloca is used to work around the lack of
deallocation on scope exit. A VLA is automatically deallocated in this
way, so it is the more fitting approach.
To implement this, it is necessary to eliminate the goto. In addition,
this change eliminates the trivially-true assert; the assert is always
skipped if nloaded > 0.
* elf/dl-fini.c (_dl_fini): Rewrite to use variable-length array
instead of extend_alloca. Change control flow to avoid a goto.
Remove assert which is trivially always true.
On hppa and ia64, the macro DL_AUTO_FUNCTION_ADDRESS() uses the
variable fptr[2] in it's own scope.
The content of fptr[] is thus undefined right after the macro exits.
Newer gcc's (>= 4.7) reuse the stack space of this variable triggering
a segmentation fault in dl-init.c:69.
To fix this we rewrite the macros to make the call directly to init
and fini without needing to pass back a constructed function pointer.
Resolves: #15465
The program name may be unavailable if the user application tampers
with argc and argv[]. Some parts of the dynamic linker caters for
this while others don't, so this patch consolidates the check and
fallback into a single macro and updates all users.
[BZ #13882]
* elf/dl-deps.c (_dl_map_object_deps): Fix cycle detection. Use
uint16_t for elements in the "seen" array to avoid char overflows.
* elf/dl-fini.c (_dl_sort_fini): Likewise.
* elf/dl-open.c (dl_open_worker): Likewise.
void * pointers instead of struct link_map **.
(_dl_scope_free): Change argument type to void *.
* include/link.h (struct link_map): Change type of l_reldeps
to struct link_map_reldeps, move l_reldepsact into that
struct too.
* elf/dl-deps.c: Include atomic.h.
(_dl_map_object_deps): Only change l->l_initfini when it is
fully populated, use _dl_scope_free for freeing it. Optimize
removal of libs from reldeps by using l_reserved flag, when
some removal is needed, allocate a new list instead of
reallocating and free the old with _dl_scope_free. Adjust
for l_reldeps and l_reldepsact changes.
* elf/dl-lookup.c (add_dependency): Likewise. Reorganize to allow
searching in l_initfini and l_reldeps without holding dl_load_lock.
* elf/dl-fini.c (_dl_sort_fini): Adjust for l_reldeps and
l_reldepsact changes.
* elf/dl-close.c (_dl_close_worker): Likewise.
* elf/dl-open.c (_dl_scope_free): Change argument type to void *.
function _dl_sort_fini.
(_dl_sort_fini): New function.
* sysdeps/generic/ldsodefs.h: Declare _dl_sort_fini.
* elf/dl-close.c (_dl_close): Call _dl_sort_fini before running
destructors to call them in the right order.
Change type of l_idx to int.
* elf/dl-close.c: Basically rewrite. Do not use l_opencount to
determine whether a DSO has to be unloaded. Instead compute this
in this function.
* elf/dl-deps.c: No need to manipulate l_opencount anymore.
* elf/dl-lookup.c: Likewise.
* elf/rtld.c: Likewise
* elf/dl-open.c: Likewise. Use l_init_called to determine whether
object was just loaded.
* elf/dl-fini.c: Bump l_direct_opencount instead of l_opencount.
* elf/dl-load.c (_dl_map_object_from_fd): Do not recognize DSO which
is about to be unloaded as a match.
(_dl_map_object): Likewise.
* elf/do-lookup.h (do_lookup_x): Do not look into DSO which is about
to be unloaded.
* elf/circleload1.c: Don't use l_opencount anymore.
* elf/neededtest.c: Likewise.
* elf/neededtest2.c: Likewise.
* elf/neededtest3.c: Likewise.
* elf/neededtest4.c: Likewise.
* elf/unload.c: Likewise.
* elf/unload2.c: Likewise.
* elf/loadtest.c: Likewise.
* elf/rtld.c: Preloading errors are now never fatal.
2005-03-08 Jakub Jelinek <jakub@redhat.com>
* elf/Makefile: Add rules to build and run unload5 test.
* elf/unload5.c: New file.
2005-03-08 Jakub Jelinek <jakub@redhat.com>
* elf/Makefile: Add rules to build and run unload4 test.
* elf/unload4.c: New file.
* elf/unload4mod1.c: New file.
* elf/unload4mod2.c: New file.
* elf/unload4mod3.c: New file.
* elf/unload4mod4.c: New file.
2004-10-27 Ulrich Drepper <drepper@redhat.com>
* elf/dl-fini.c (_dl_fini): Fix search for map in maps array.
Reverse order of namespaces.
* elf/Makefile: Add rules to build and run tst-dlmopen3.
* elf/tst-dlmopen3.c: New file.
* elf/tst-dlmopen1mod.c: Add check whether constructor runs.
Update.
Add support for namespaces in the dynamic linker.
* dlfcn/Makefile (libdl-routines): Add dlmopen.
* dlfcn/Versions [libdl, GLIBC_2.3.4]: Add dlmopen.
* dlfcn/dlfcn.h: Define Lmid_t, LM_ID_BASE, and LM_ID_NEWLM.
Declare dlmopen. Document RTLD_DI_LMID.
* dlfcn/dlinfo.c: Handle RTLD_DI_LMID.
* dlfcn/dlmopen.c: New file.
* dlfcn/dlopen.c: Pass new parameter to _dl_open.
* dlfcn/dlopenold.c: Likewise.
* elf/dl-addr.c: Adjust for removal of GL(dl_loaded).
* elf/dl-caller.c: Likewise.
* elf/dl-close.c: Likewise.
* elf/dl-conflict.c: Likewise.
* elf/dl-debug.c: Likewise.
* elf/dl-lookup.c: Likewise.
* elf/dl-sym.c: Likewise.
* elf/dl-version.c: Likewise.
* elf/do-lookup.h: Likewise.
* elf/rtld.c: Likewise.
* sysdeps/unix/sysv/linux/i386/dl-librecon.h: Likewise.
* elf/dl-depsc: Likewise. Add new parameter to _dl_map_object.
* elf/dl-fini.c: Call destructors in all namespaces.
* elf/dl-iteratephdr.c: Compute total nloaded. Adjust for removal of
GL(dl_loaded).
* elf/dl-libc.c: Pass new parameter to _dl_open. Adjust for removal
of GL(dl_loaded).
* elf/dl-load.c (_dl_map_object_from_fd): Don't load ld.so a second
time. Reuse the one from the main namespace in all others.
Pass new parameter to _dl_new_object.
Adjust for removal of GL(dl_loaded).
* elf/dl-object.c: Take new parameter. Use it to initialize l_ns.
Adjust for removal of GL(dl_loaded).
* elf/dl-open.c (_dl_open): Take new parameter.
Adjust for removal of GL(dl_loaded).
* elf/dl-support.c: Replace global _dl_loaded etc variables with
_dl_ns variable.
* include/dlfcn.h: Adjust prototype of _dl_open.
Define __LM_ID_CALLER.
* include/link.h: Add l_real, l_ns, and l_direct_opencount elements.
* sysdeps/generic/dl-tls.c: Bump TLS_STATIC_SURPLUS. Since libc is
using TLS we need memory appropriate to the number of namespaces.
* sysdeps/generic/ldsodefs.h (struct rtld_global): Replace _dl_loaded,
_dl_nloaded, _dl_global_scope, _dl_main_searchlist, and
_dl_global_scope_alloc with _dl_ns element. Define DL_NNS.
Adjust prototypes of _dl_map_object and member in rtld_global_ro.
* malloc/malloc.c: Include <dlfcn.h>.
* malloc/arena.c (ptmalloc_init): If libc is not in primary namespace,
never use brk.
* elf/Makefile: Add rules to build and run tst-dlmopen1 and
tst-dlmopen2.
* elf/tst-dlmopen1.c: New file.
* elf/tst-dlmopen1mod.c: New file.
* elf/tst-dlmopen2.c: New file.
* elf/dl-close.c: Improve reference counting by tracking direct loads.
* elf/dl-lookup.c (add_dependency): Likewise.
* elf/dl-open.c (dl_open_worker): Likewise.
* elf/rtld.c (dl_main): Likewise.
2004-09-09 GOTO Masanori <gotom@debian.or.jp>
[BZ #77]
* elf/dl-close.c: Count down l_opencount to check not only for
l_reldeps, but also l_initfini.
2004-10-13 Ulrich Drepper <drepper@redhat.com>
