From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from zeniv.linux.org.uk ([195.92.253.2]:46899 "EHLO ZenIV.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751055AbcBARfD (ORCPT ); Mon, 1 Feb 2016 12:35:03 -0500 Date: Mon, 1 Feb 2016 17:35:01 +0000 From: Al Viro To: David Howells Cc: linux-fsdevel@vger.kernel.org, linux-nfs@vger.kernel.org, linux-cachefs@redhat.com Subject: Re: [PATCH] CacheFiles: Provide read-and-reset release counters for cachefilesd Message-ID: <20160201173501.GL17997@ZenIV.linux.org.uk> References: <20160201164304.9867.40456.stgit@warthog.procyon.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20160201164304.9867.40456.stgit@warthog.procyon.org.uk> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Mon, Feb 01, 2016 at 04:43:04PM +0000, David Howells wrote: > Provide read-and-reset objects- and blocks-released counters for cachefilesd > to use to work out whether there's anything new that can be culled. > > One of the problems cachefilesd has is that if all the objects in the cache > are pinned by inodes lying dormant in the kernel inode cache, there isn't > anything for it to cull. In such a case, it just spins around walking the > filesystem tree and scanning for something to cull. This eats up a lot of > CPU time. Applied.