Date:   Fri, 25 Jan 2019 15:30:01 -0500
From:   Bruce Fields <bfields@fieldses.org>
To:     Chuck Lever <chuck.lever@oracle.com>
Cc:     Trond Myklebust <trondmy@hammerspace.com>,
        Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH] SUNRPC: Don't allow compiler optimisation of
 svc_xprt_release_slot()
Message-ID: <20190125203001.GA5972@fieldses.org>
References: <4077991d3d3acee4c37c7c8c6dc2b76930c9584e.camel@hammerspace.com>
 <20190109165142.GB32189@fieldses.org>
 <300445038b75d5efafe9391eb4b8e83d9d6e3633.camel@hammerspace.com>
 <20190111211235.GA27206@fieldses.org>
 <BB7240BE-3F20-43AD-B709-7F18CC33CC02@oracle.com>
 <6F5B73B7-E9F8-4FDB-8381-E5C02772C6A5@oracle.com>
 <20190111221030.GA28794@fieldses.org>
 <A33F8CE5-3E5D-4359-8061-8DB8214B012B@oracle.com>
 <20190112005613.GA29181@fieldses.org>
 <F21839E5-A5DB-4BC8-8F3B-2AD02D4B2652@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <F21839E5-A5DB-4BC8-8F3B-2AD02D4B2652@oracle.com>
User-Agent: Mutt/1.5.21 (2010-09-15)
Sender: linux-nfs-owner@vger.kernel.org
Precedence: bulk

On Mon, Jan 14, 2019 at 12:24:24PM -0500, Chuck Lever wrote:
> 
> 
> > On Jan 11, 2019, at 7:56 PM, Bruce Fields <bfields@fieldses.org> wrote:
> > 
> > On Fri, Jan 11, 2019 at 05:27:30PM -0500, Chuck Lever wrote:
> >> 
> >> 
> >>> On Jan 11, 2019, at 5:10 PM, Bruce Fields <bfields@fieldses.org> wrote:
> >>> 
> >>> On Fri, Jan 11, 2019 at 04:54:01PM -0500, Chuck Lever wrote:
> >>>>> On Jan 11, 2019, at 4:52 PM, Chuck Lever <chuck.lever@oracle.com> wrote:
> >>>>>> So, I think we need your patch plus something like this.
> >>>>>> 
> >>>>>> Chuck, maybe you could help me with the "XXX: Chuck:" parts?
> >>>>> 
> >>>>> I haven't been following. Why do you think those are necessary?
> >>> 
> >>> I'm worried something like this could happen:
> >>> 
> >>> 	CPU 1				CPU 2
> >>> 	-----				-----
> >>> 
> >>> 	set XPT_DATA			dec xpt_nr_rqsts
> >>> 
> >>> 	svc_xprt_enqueue		svc_xprt_enqueue
> >>> 
> >>> And both decide nothing should be done if neither sees the change that
> >>> the other made.
> >>> 
> >>> Maybe I'm still missing some reason that couldn't happen.
> >>> 
> >>> Even if it can happen, it's an unlikely race that will likely be fixed
> >>> when another event comes along a little later, which would explain why
> >>> we've never seen any reports.
> >>> 
> >>>>> We've had set_bit and atomic_{inc,dec} in this code for ages,
> >>>>> and I've never noticed a problem.
> >>>>> 
> >>>>> Rather than adding another CPU pipeline bubble in the RDMA code,
> >>>>> though, could you simply move the set_bit() call site inside the
> >>>>> critical sections?
> >>>> 
> >>>> er, inside the preceding critical section. Just reverse the order
> >>>> of the spin_unlock and the set_bit.
> >>> 
> >>> That'd do it, thanks!
> >> 
> >> I can try that here and see if it results in a performance regression.
> > 
> > Thanks, I've got a version with a typo fixed at
> > 
> > 	git://linux-nfs.org/~bfields/linux.git nfsd-next
> 
> Applied all four patches here. I don't see any performance regressions,
> but my server has only a single last-level CPU cache.

Thanks!

I'm adding a Tested-by: for you if that's OK.

--b.