[omniORB] Thread omniORB synchroniztion bug on startup?

[Sai-Lai Lo]

>>>>> Rob Kovalchik writes:

> A very simple client running on Solaris 2.6 with omniORB 2.6.0 will
> periodically crash during startup, prior to entering "main()" with a SE=
GV.
> This occurs in omni_thread::init_t::init_t() as shown by the following =
stack
> trace from the resulting core file using dbx.  The client app usually w=
orks
> fine. This error occurs maybe one out of ten startups.

We have not experienced the problem you described. The compiler we are
using is:

% CC -V
CC: WorkShop Compilers 4.2 30 Oct 1996 C++ 4.2

Don't know if this is relevant.

=46rom you stack trace, it looks like an exception has been raised inside=

omni_thread::init_t::init_t(). This function initialised the omnithread
library and would raise an exception if any calls into the pthread librar=
y
fails. The function is called as a side effect of the global initialiser
for the variable omni_thread_init. Because it is a global initialiser, it=

is called before main is called. Could there be a bug in the
linker/compiler that causes omni_thread_init to be called before the
pthread library is initialised?

Regards,
Sai-Lai



-- =

Dr. Sai-Lai Lo                          |       Research Scientist
                                        |
E-mail:         S.Lo@orl.co.uk          |       Olivetti & Oracle Researc=
h Lab
                                        |       24a Trumpington Street
Tel:            +44 223 343000          |       Cambridge CB2 1QA
Fax:            +44 223 313542          |       ENGLAND