Lazy binding failures during dlopen/dlclose must be fatal [BZ #24304]

If a lazy binding failure happens during the execution of an ELF
constructor or destructor, the dynamic loader catches the error
and reports it using the dlerror mechanism.  This is undesirable
because there could be other constructors and destructors that
need processing (which are skipped), and the process is in an
inconsistent state at this point.  Therefore, we have to issue
a fatal dynamic loader error error and terminate the process.

Note that the _dl_catch_exception in _dl_open is just an inner catch,
to roll back some state locally.  If called from dlopen, there is
still an outer catch, which is why calling _dl_init via call_dl_init
and a no-exception is required and cannot be avoiding by moving the
_dl_init call directly into _dl_open.

_dl_fini does not need changes because it does not install an error
handler, so errors are already fatal there.

Change-Id: I6b1addfe2e30f50a1781595f046f44173db9491a

1 files changed, 6 insertions, 0 deletions

diff --git a/NEWS b/NEWS
index 12b239c..5a2d0a5 100644
--- a/NEWS
+++ b/NEWS
@@ -93,6 +93,12 @@ Deprecated and removed features, and other changes affecting compatibility:
   are no longer supported.  For v8 only implementations with native CAS
   instruction are still supported (such as LEON).
 
+* If a lazy binding failure happens during dlopen, during the execution of
+  an ELF constructor, the process is now terminated.  Previously, the
+  dynamic loader would return NULL from dlopen, with the lazy binding error
+  captured in a dlerror message.  In general, this is unsafe because
+  resetting the stack in an arbitrary function call is not possible.
+
 Changes to build and runtime requirements:
 
   [Add changes to build and runtime requirements here]