dlopen() and dlclose() are not MT-safe?

[Kazuaki Oda]

Kostik Belousov wrote:
> Oops. Completely reversed condition in the if. :(. Also, I don't think it
> shall returns the error in this situation. New take:
> 
> Index: libexec/rtld-elf/rtld.c
> ===================================================================
> RCS file: /usr/local/arch/ncvs/src/libexec/rtld-elf/rtld.c,v
> retrieving revision 1.112
> diff -u -r1.112 rtld.c
> --- libexec/rtld-elf/rtld.c     24 Dec 2005 15:37:30 -0000      1.112
> +++ libexec/rtld-elf/rtld.c     22 Mar 2006 19:03:12 -0000
> @@ -1688,6 +1688,11 @@
>         wlock_release(rtld_bind_lock, lockstate);
>         objlist_call_fini(&list_fini);
>         lockstate = wlock_acquire(rtld_bind_lock);
> +       if (root->refcount != 0) {
> +               wlock_release(rtld_bind_lock, lockstate);
> +               return 0;
> +       }
> +
>         objlist_remove_unref(&list_fini);
> 
>         /* Finish cleaning up the newly-unreferenced objects. */

Thanks.  I applied your patch and ran the test program once more.

When linked with libpthread:
% ./dltest
Segmentation fault (core dumped)

When linked with libthr:
% ./dltest
ld-elf.so.1: assert failed: /usr/src/libexec/rtld-elf/rtld.c:1728

In libpthread case, the result was changed.  I got no "assert failed"
message, and it took longer time to crash than before.  In libthr
case, I could not find any difference.

BTW do you know the reason why lock is released before calling
objlist_call_fini()?  If we don't release the lock, what problem will
occur?  deadlock?

--
Kazuaki Oda