Re: performance delta after VFS i_mutex=>i_rwsem conversion

From: Ingo Molnar <mingo@kernel.org>
To: Waiman Long <waiman.long@hpe.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Dave Hansen <dave.hansen@intel.com>,
	"Chen, Tim C" <tim.c.chen@intel.com>,
	Ingo Molnar <mingo@redhat.com>, Davidlohr Bueso <dbueso@suse.de>,
	"Peter Zijlstra (Intel)" <peterz@infradead.org>,
	Jason Low <jason.low2@hp.com>,
	Michel Lespinasse <walken@google.com>,
	"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
	Waiman Long <waiman.long@hp.com>,
	Al Viro <viro@zeniv.linux.org.uk>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: performance delta after VFS i_mutex=>i_rwsem conversion
Date: Wed, 8 Jun 2016 10:58:37 +0200	[thread overview]
Message-ID: <20160608085837.GA10792@gmail.com> (raw)
In-Reply-To: <5755E782.90800@hpe.com>

* Waiman Long <waiman.long@hpe.com> wrote:

> I do have a patchset that allow us to more accurately determine the state of
> the lock owner.
> 
> locking/rwsem: Add reader-owned state to the owner field
> http://www.spinics.net/lists/kernel/msg2258572.html
> 
> That should eliminate the performance gap between mutex and rwsem wrt
> spinning when only writers are present. I am hoping that that patchset can
> be queued for 4.8.

Yeah, so I actually had this series merged for testing last week, but a 
complication with a prereq patch made me unmerge it. But I have no fundamental 
objections, at all.

I also agree with Linus's general observation that we want to make 
down_write()/up_write() match mutex performance characteristics.

I think kernel developers should fundamentally be able to switch between 
mutex_lock()/unlock() and down_write()/up_write() and back, without noticing
any high level behavioral changes.

Any 'reader/writer mixing' artifacts are secondary concerns and we'll sort them 
out as they happen.

Thanks,

	Ingo