[omniORB] Significance of --enable-threads for building egcs-1.1

[Sai-Lai Lo]

Shinji,

I didn't read the example carefully and you are right the process limit is
not the problem.

I ran your test and got some interesting result.

I can reproduce your problem on a Redhat 5.1 machine with a 2.0.34 kernel.

*BUT* when I repeated the test on a machine with a 2.1.125 kernel, the
 problem goes away.

Seems to me its a bug in the kernel which has since been fixed in 2.1.

Hope you can reproduce my result.

Sai-Lai



>>>>> Shinji Suzuki writes:

> Encouraged by your question, I put exception handlers around
> the relevant code fragment and found that, after 1022 creations,
> omni_thread_invalid(11) is thrown from 'omni_thread::create(here,0)'.
> I wish I could have investigated futher immediately now but setting
> up multi-thread debugging environment (building new gdb and
> patching/building glibc) is lots of work to say the least. Anyway,
> time to have a look at omnithread source code.




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