From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from zeniv.linux.org.uk ([195.92.253.2]:35672 "EHLO ZenIV.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754181AbcKKPkh (ORCPT ); Fri, 11 Nov 2016 10:40:37 -0500 Date: Fri, 11 Nov 2016 15:40:34 +0000 From: Al Viro To: Amir Goldstein Cc: Miklos Szeredi , Konstantin Khlebnikov , linux-unionfs@vger.kernel.org, linux-fsdevel Subject: Re: [RFC][PATH 4/4] ovl: relax lock_rename when moving files between work and upper dir Message-ID: <20161111154034.GR19539@ZenIV.linux.org.uk> References: <1478817883-27662-5-git-send-email-amir73il@gmail.com> <20161110230221.GL19539@ZenIV.linux.org.uk> <20161110230557.GM19539@ZenIV.linux.org.uk> <20161110231757.GN19539@ZenIV.linux.org.uk> <20161110235444.GO19539@ZenIV.linux.org.uk> <20161111002756.GP19539@ZenIV.linux.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Fri, Nov 11, 2016 at 04:43:57PM +0200, Amir Goldstein wrote: > > Surely, the copying of data itself is outside of that lock, isn't it? > > Mmmmmm, no it isn't, but I am going to make it right. Well, if you really have the copyup itself done under lock_rename(), you are looking at potentially copying gigabytes of data from r/o layer to workdir under fs-wide lock. _That_ is really asking for contention from hell, and for no good reason. Renames/whiteout creations/removals are trivial noise compared to that. I'm somewhat surprised, TBH - I thought the copyup itself is done outside of that thing. Are you sure it really isn't?