On Mon, May 26, 2008 at 09:34:41PM -0700, Daniel Walker wrote:
quoted text
> 
> On Mon, 2008-05-26 at 20:15 -0600, Matthew Wilcox wrote:
> > On Tue, May 27, 2008 at 11:17:22AM +1000, Rusty Russell wrote:
> > > On Tuesday 27 May 2008 03:29:14 Michal Schmidt wrote:
> > > > Mutexes are not allowed in interrupt context, not even mutex_trylock.
> > > 
> > > As an aside, does anyone know why?  I know the documentation says so, but it 
> > > wasn't immediately obvious to me.  I asked before to no response...
> > 
> > Because mutexes have an owner.  In interrupt context, there is no owner.
> > This owner is used to do priority boosts as well as debugging.
> 
> I don't think regular mutexes do boosting, the rtmutex does tho .. I
> think it's more a issue with lockdep complaining about the context, but
> that exists to just plain disallow this type of usage on principal .

You're right, I got confused, it's only rtmutexes that do priority
inheritance.

I suppose we could have a special 'interrupt context' owner for
the purposes of LOCKDEP.  Ingo?

-- 
Intel are signing my paycheques ... these opinions are still mine
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/