From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pl1-f198.google.com (mail-pl1-f198.google.com [209.85.214.198]) by kanga.kvack.org (Postfix) with ESMTP id 3043B6B487F for ; Tue, 28 Aug 2018 18:50:09 -0400 (EDT) Received: by mail-pl1-f198.google.com with SMTP id w18-v6so1272688plp.3 for ; Tue, 28 Aug 2018 15:50:09 -0700 (PDT) Received: from mail.linuxfoundation.org (mail.linuxfoundation.org. [140.211.169.12]) by mx.google.com with ESMTPS id s9-v6si2430913pgn.286.2018.08.28.15.50.07 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Tue, 28 Aug 2018 15:50:08 -0700 (PDT) Date: Tue, 28 Aug 2018 15:50:06 -0700 From: Andrew Morton Subject: Re: [PATCH 0/2] fs/dcache: Track # of negative dentries Message-Id: <20180828155006.a6a94a7ba64ac4ce6b8b190c@linux-foundation.org> In-Reply-To: <1535476780-5773-1-git-send-email-longman@redhat.com> References: <1535476780-5773-1-git-send-email-longman@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-ID: To: Waiman Long Cc: Alexander Viro , Jonathan Corbet , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-mm@kvack.org, linux-doc@vger.kernel.org, "Luis R. Rodriguez" , Kees Cook , Linus Torvalds , Jan Kara , "Paul E. McKenney" , Ingo Molnar , Miklos Szeredi , Matthew Wilcox , Larry Woodman , James Bottomley , "Wangkai (Kevin C)" , Michal Hocko On Tue, 28 Aug 2018 13:19:38 -0400 Waiman Long wrote: > This patchset is a reduced scope version of the > patchset "fs/dcache: Track & limit # of negative dentries" > (https://lkml.org/lkml/2018/7/12/586). Only the first 2 patches are > included to track the number of negative dentries in the system as well > as making negative dentries more easily reclaimed than positive ones. > > There are controversies on limiting number of negative dentries as it may > make negative dentries special in term of how memory resources are to > be managed in the kernel. However, I don't believe I heard any concern > about tracking the number of negative dentries in the system. So it is > better to separate that out and get it done with. We can deal with the > controversial part later on. Seems reasonable. It would be nice to see testing results please. Quite comprehensive ones. And again, an apparently permanent feature of this patchset is that the changelogs fail to provide descriptions of real-world problems with the existing code. Please do provide those (comprehensive) descriptions and demonstrate that these changes resolve those problems. Also, a grumpynit: with 100% uniformity, the vfs presently refers to negative dentries with the string "negative" in the identifier. This patchset abbreviates that to "neg".