Hello,
I have a problem when starting SER 0.9.0 an a linux ppc64. The trace is as follow:
... unixsock_register_cmd: New command (t_uac_cancel) registered 0(0) unixsock_register_cmd: New command (t_hash) registered 0(0) unixsock_register_cmd: New command (t_reply) registered 0(0) DEBUG: lock_initialize: lock initialization started 0(0) ERROR: lock_set_init (SYSV): semget (..., 251, 0700) failed: Invalid argument 0(0) DEBUG: lock_initialize: entry semaphore initialization failure: Invalid argument 0(0) DEBUG: lock_initialize: first time semaphore allocation failure 0(0) INFO: semaphore arrays of size 250 allocated ...
This error was already reported on april, 2005 and Jan Janak suggested changing SEMMSL value in include files. I did that, but the error continues...
Do someone know something about this problem?
Thank you!
On May 20, 2005 at 16:15, Jorge Crichigno jcrichigno@conexion.com.py wrote:
Hello,
I have a problem when starting SER 0.9.0 an a linux ppc64. The trace is as follow:
... unixsock_register_cmd: New command (t_uac_cancel) registered 0(0) unixsock_register_cmd: New command (t_hash) registered 0(0) unixsock_register_cmd: New command (t_reply) registered 0(0) DEBUG: lock_initialize: lock initialization started 0(0) ERROR: lock_set_init (SYSV): semget (..., 251, 0700) failed: Invalid argument 0(0) DEBUG: lock_initialize: entry semaphore initialization failure: Invalid argument 0(0) DEBUG: lock_initialize: first time semaphore allocation failure 0(0) INFO: semaphore arrays of size 250 allocated
Don't worry, this is normal, it happens when ser probes for the maximum size of the semaphore array.
See http://lists.iptel.org/pipermail/serdev/2005-April/004339.html.
On the other hand, ser would work much better with native ppc64 locking support (but for this I would need access to a ppc64 machine, or someone willing to compile some test programs and report back errors).
Could you send the uname -m output?
Andrei
-
Andrei Pelinescu-Onciul -
Jorge Crichigno