Re: mutex ownership (was: Re: [PATCH 19/24] Unionfs: Helper macros/inlines)

[Steven Rostedt <rostedt@goodmis.org>]

On Tue, 2007-01-09 at 10:09 +0100, Oliver Neukum wrote:
> Am Dienstag, 9. Januar 2007 10:02 schrieb Peter Zijlstra:

> > Its a fundamental property of a mutex, not a shortcoming. A mutex has an
> > owner, the one that takes and releases the resource. This allows things
> > such as Priority Inheritance to boost owners.
> > 
> > 'fixing' this takes away much of what a mutex is.
> > 
> > That said, it seems some folks really want this to happen, weird as it
> > may be. I'm not sure if all these cases are because of wrong designs. A
> > possible extension to the mutex interface might be something like this:
> > 
> >   mutex_pass_owner(struct task_struct *task);
> > 
> > which would be an atomic unlock/lock pair where the current task
> > releases the resource and the indicated task gains it. However it must
> > be understood that from the POV of 'current' this should be treated as
> > an unlock action.
> 
> This won't help if I want to release from an interrupt handler or tasklet.

Then that shouldn't use a mutex. A mutex is for "mutual exclusion", and
not "shared exclusion".  Peter is right that the fundamental property of
a mutex, is that it is for a single thread.  And it should never be
released by a separate thread than the one that grabbed it.

A semaphore can be used as a mutex, but not the other way around.

I wouldn't even recommend doing the mutex_pass_owner().  To do so, the
process calling this must be the owner, and the parameter must be a
thread waiting on the mutex. If it is not, then that could introduce
some really nasty races and bugs.

The mutex_unlock_dont_warn_if_a_different_task_did_it() is also a bad
idea.

-- Steve