Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 9 Nov 2020 15:20:16 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     Trond Myklebust <trondmy@hammerspace.com>
Cc:     "neilb@suse.de" <neilb@suse.de>,
        "juri.lelli@redhat.com" <juri.lelli@redhat.com>,
        "mingo@redhat.com" <mingo@redhat.com>,
        "jiangshanlai@gmail.com" <jiangshanlai@gmail.com>,
        "tj@kernel.org" <tj@kernel.org>,
        "mhocko@suse.com" <mhocko@suse.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "vincent.guittot@linaro.org" <vincent.guittot@linaro.org>
Subject: Re: [PATCH rfc] workqueue: honour cond_resched() more effectively.
Message-ID: <20201109142016.GK2611@hirez.programming.kicks-ass.net>
References: <87v9efp7cs.fsf@notabene.neil.brown.name>
 <20201109080038.GY2594@hirez.programming.kicks-ass.net>
 <aec65c71c09e803285688d5974193a98b4422428.camel@hammerspace.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <aec65c71c09e803285688d5974193a98b4422428.camel@hammerspace.com>
Precedence: bulk

On Mon, Nov 09, 2020 at 01:50:40PM +0000, Trond Myklebust wrote:
> On Mon, 2020-11-09 at 09:00 +0100, Peter Zijlstra wrote:

> > I'm thinking the real problem is that you're abusing workqueues. Just
> > don't stuff so much work into it that this becomes a problem. Or
> > rather,
> > if you do, don't lie to it about it.
> 
> If we can't use workqueues to call iput_final() on an inode, then what
> is the point of having them at all?

Running short stuff, apparently.

> Neil's use case is simply a file that has managed to accumulate a
> seriously large page cache, and is therefore taking a long time to
> complete the call to truncate_inode_pages_final(). Are you saying we
> have to allocate a dedicated thread for every case where this happens?

I'm not saying anything, but you're trying to wreck the scheduler
because of a workqueue 'feature'. The 'new' workqueues limit concurrency
by design, if you're then relying on concurrency for things, you're
using it wrong.

I really don't know what the right answer is here, but I thoroughly hate
the one proposed.