Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Thu, 18 Oct 2012 10:39:21 -0500

Hi Ralph,

On 2012-10-18 Ralph Lange wrote:
> > BTW, if pthread_spin_lock() is not implemented, using a native
> > pthread_mutex_t might be lighter weight than an epicsMutex object.
> 
> On Posix, yes. (Only slightly though, as the Posix implementation of
> epicsMutex uses pthread_mutex_t.)

The *second* implementation in osi/posix/osdMutex.c (for Posix OSs where 
PTHREAD_MUTEX_RECURSIVE is  not available) is where it will make more of a 
difference, and that is more likely to be used where pthread_spin_lock() is 
not available.  Not as important though I agree.

> The Default should be epicsMutex, to make things work on Windows etc.
> until specific implementations are in place.

At the moment most target builds (other than self-hosted Windows) search the 
os/posix source directories for files before os/default, so you can't rely on 
the os/default implementation being used if there is an os/posix one.  This is 
a bit stupid, but I haven't tried setting POSIX=NO in the vxWorks or RTEMS 
target configuration files so I don't know how hard that might be to change.

Thanks,

- Andrew
-- 
Never interrupt your enemy when he is making a mistake.
-- Napoleon Bonaparte

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System