Re: [RFC][PATCH 5/8] rtmutex: Clean up

From	Juri Lelli <juri.lelli@arm.com>
Newsgroups	linux.kernel
Subject	Re: [RFC][PATCH 5/8] rtmutex: Clean up
Date	2016-06-14 14:50 +0200
Message-ID	<rJVX3-4SP-17@gated-at.bofh.it> (permalink)
References	<rHvDI-7fg-35@gated-at.bofh.it> <rHvNo-7ix-25@gated-at.bofh.it> <rJVkl-4Ei-13@gated-at.bofh.it> <rJVNn-4Px-9@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

On 14/06/16 14:32, Peter Zijlstra wrote:
> On Tue, Jun 14, 2016 at 01:08:13PM +0100, Juri Lelli wrote:
> > > +	postunlock = rt_mutex_futex_unlock(&pi_state->pi_mutex, &wake_q);
> > >  
> > >  	/*
> > >  	 * First unlock HB so the waiter does not spin on it once he got woken
> > > +	 * up. Then wakeup the waiter by calling rt_mutex_postunlock(). Priority
> > > +	 * is already adjusted and preemption is disabled to avoid inversion.
> > >  	 */
> > >  	spin_unlock(&hb->lock);
> > >  
> > > +	if (postunlock)
> > > +		rt_mutex_postunlock(&wake_q);
> > 
> > I'm most probably missing something, but don't we still need to call
> > wake_up_q() even when postunlock is false? IIUC, we were always doing
> > that, rt_mutex_postunlock(), before this change (only calling
> > preempt_enable() was conditional).
> 
> Note that rt_mutex_slowunlock() only uses wake_q on the true path. When
> it returns false, it will not have placed a task to wake up.
> 

Right. But, I thought we were still ending up calling wake_up_q before
this change. Which however looked fine, as it won't do anything if no
task is queued I guess. So, no problem before and no problem now I'd
say. :-)

Back to linux.kernel | Previous | Next — Previous in thread | Find similar | Unroll thread

Thread

Re: [RFC][PATCH 5/8] rtmutex: Clean up Juri Lelli <juri.lelli@arm.com> - 2016-06-14 14:10 +0200
  Re: [RFC][PATCH 5/8] rtmutex: Clean up Peter Zijlstra <peterz@infradead.org> - 2016-06-14 14:40 +0200
    Re: [RFC][PATCH 5/8] rtmutex: Clean up Juri Lelli <juri.lelli@arm.com> - 2016-06-14 14:50 +0200

csiph-web