Return-Path: linux-nfs-owner@vger.kernel.org Received: from bombadil.infradead.org ([198.137.202.9]:44732 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932556Ab3LDRGF (ORCPT ); Wed, 4 Dec 2013 12:06:05 -0500 Date: Wed, 4 Dec 2013 09:06:02 -0800 From: Christoph Hellwig To: Jeff Layton Cc: "J. Bruce Fields" , Christoph Hellwig , gartim@gmail.com, linux-nfs@vger.kernel.org Subject: Re: [PATCH] nfsd: when reusing an existing repcache entry, unhash it first Message-ID: <20131204170602.GA13308@infradead.org> References: <1386015979-27511-1-git-send-email-jlayton@redhat.com> <20131203102517.GA12576@infradead.org> <20131203132112.1f19c014@tlielax.poochiereds.net> <20131204083336.GB30216@infradead.org> <20131204075402.7b00d09d@tlielax.poochiereds.net> <20131204130944.GA3464@infradead.org> <20131204142509.GA14646@fieldses.org> <20131204120616.464e79d8@corrin.poochiereds.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20131204120616.464e79d8@corrin.poochiereds.net> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Wed, Dec 04, 2013 at 12:06:16PM -0500, Jeff Layton wrote: > That is another tricky question. Currently, we size the number of hash > buckets based on the max size of the cache. If we allow it to grow > without bound, then how do we determine how many hash buckets we need? It won't grow without bound as long as it's trimmed below the size by both the periodic workqueue and the shrinker. Of course replacing the hash with an autoscaling data structure would be useful, but given that 4.1 obsoletes the DRC I'm not sure how much effort we want to put into it.