Experimental Physics and
Industrial Control System

Marty Kraimer <[email protected]> · Tue, 03 Dec 2002 14:14:14 -0600

I have been looking at Jeff's suggestion about implementing a faster version of

osi/os/posix/osdMutex.c. The idea is to use PTHREAD_MUTEX_RECURSIVE if posix

implementation allows it.

One catch is epicsMutexLockWithTimeout. There is no pthread_mutux_lock with a

timeout.

I just got done talking on the phone with Jeff. We both think the solution is to

get rid of epicsMutexLockWithTimeout. It is only used in two places in base and

both are in diagnostic routines.

Unless I hear major objections epicsMutexLockWithTimeout will be removed.

Marty