Re: [PATCH 2/4] locking/rtmutex: Use cmp-cmpxchg

[Jason Low <jason.low2@hp.com>]

Hi David,

On Sat, Jun 6, 2015 at 8:27 AM, Davidlohr Bueso <dave@stgolabs.net> wrote:
> On Fri, 2015-06-05 at 14:38 +0200, Thomas Gleixner wrote:
>> On Tue, 19 May 2015, Davidlohr Bueso wrote:
>>
>> > Avoid unnecessary cmpxchg calls, all of our other locks
>> > use it as well.
>> >
>> > Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
>> > ---
>> >  kernel/locking/rtmutex.c | 4 +++-
>> >  1 file changed, 3 insertions(+), 1 deletion(-)
>> >
>> > diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
>> > index 74188d8..1d5cc9d 100644
>> > --- a/kernel/locking/rtmutex.c
>> > +++ b/kernel/locking/rtmutex.c
>> > @@ -74,7 +74,9 @@ static void fixup_rt_mutex_waiters(struct rt_mutex *lock)
>> >   * set up.
>> >   */
>> >  #ifndef CONFIG_DEBUG_RT_MUTEXES
>> > -# define rt_mutex_cmpxchg(l,c,n)   (cmpxchg(&l->owner, c, n) == c)
>> > +# define rt_mutex_cmpxchg(l,c,n) \
>> > +   (l->owner == c && cmpxchg(&l->owner, c, n) == c)
>>
>> Errm. How does that improve stuff like rt_mutex_lock() ?
>
> It just avoids a cmpxchg in the fastpath when locked, at the cost of an
> extra test when unlocked. CCAS techniques have been proven to boost some
> workloads for both rwsems and mutexes. That's all.

The CCAS technique was typically used in the slow paths for those
other locks, where the chance of the operation returning false is
higher.

The rt_mutex_cmpxchg is used in places such as rt_mutex fastlocks. We
might not want to add extra costs to the fast paths, particularly when
the rt_mutex_cmpxchg are marked "likely" in those cases.