Near Last Call Before Branching 4.11

Peter Dufault dufault at hda.com
Thu Jul 16 20:30:23 UTC 2015


> On Jul 16, 2015, at 15:07 , Joel Sherrill <Joel.Sherrill at oarcorp.com> wrote:
> 
>> 
>> - I was having NFS crashes regularly.  I removed a pthread_mutex_trylock() in an ISR (I assumed that was safe) while trying to make that ISR faster and the NFS issues disappeared.  I haven’t had time to go back to that either.
> 
> There are checks against that. I don't remember the exact restrictions and
> it may be completely forbidden. But depending on the configuration of the
> mutex, it doesn't even make sense. Priority inheritance and ceiling imply
> an owner thread so lock operations need to be from threads for mutexes.
> 
> You probably could have gotten by with a semaphore but wouldn't have had
> the nice properties of a mutex.
> 
> 

I was foolishly thinking any non-blocking routine usable in a signal handler should be usable in an ISR.  The way I’m using it can and should be replaced by a compare-and-swap.  I hope it was the cause of the NFS issues, when I have time I’m going to re-enable it and see if the problems come back.

Peter
-----------------
Peter Dufault
HD Associates, Inc.      Software and System Engineering




More information about the devel mailing list