Re: [PATCH] percpu_counter: return precise count from __percpu_counter_compare()

From: Tejun Heo <tj@kernel.org>
To: Dave Chinner <david@fromorbit.com>
Cc: Scott J Norton <scott.norton@hpe.com>,
	Christoph Lameter <cl@linux-foundation.org>,
	Douglas Hatch <doug.hatch@hpe.com>,
	linux-kernel@vger.kernel.org, Waiman Long <waiman.long@hpe.com>,
	xfs@oss.sgi.com
Subject: Re: [PATCH] percpu_counter: return precise count from __percpu_counter_compare()
Date: Wed, 7 Oct 2015 16:20:10 -0700	[thread overview]
Message-ID: <20151007232010.GA21142@mtj.duckdns.org> (raw)
In-Reply-To: <20151007230441.GG32150@dastard>

Hello, Dave.

On Thu, Oct 08, 2015 at 10:04:42AM +1100, Dave Chinner wrote:
...
> As it is, the update race you pointed out is easy to solve with
> __this_cpu_cmpxchg rather than _this_cpu_sub (similar to mod_state()
> in the MM percpu counter stats code, perhaps).

percpu cmpxchg is no different from sub or any other operations
regarding cross-CPU synchronization.  They're safe iff the operations
are on the local CPU.  They have to be made atomics if they need to be
manipulated from remote CPUs.

That said, while we can't manipulate the percpu counters directly, we
can add a separate global counter to cache sum result from the
previous run which gets automatically invalidated when any percpu
counter overflows.  That should give better and in case of
back-to-back invocations pretty good precision compared to just
returning the global overflow counter.  Interface-wise, that'd be a
lot easier to deal with although I have no idea whether it'd fit this
particular use case or whether this use case even exists.

Thanks.

-- 
tejun

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs