From mboxrd@z Thu Jan 1 00:00:00 1970 From: "J. R. Okajima" Subject: Re: [Patch] Support overriding uid/gid in overlayfs Date: Fri, 01 Feb 2013 13:45:34 +0900 Message-ID: <7500.1359693934@jrobl> References: <1358254695.3380.16.camel@avalon> <1359388685.2499.16.camel@avalon> <4997.1359651713@jrobl> Cc: Alessandro Pignotti , Linux-Fsdevel To: Miklos Szeredi Return-path: Received: from mail02-md.ns.itscom.net ([175.177.155.112]:50251 "EHLO mail02-md.ns.itscom.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752632Ab3BAEpj (ORCPT ); Thu, 31 Jan 2013 23:45:39 -0500 In-Reply-To: Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Miklos Szeredi: > They don't "solve" this, it's a fundamental property of the > implementation. Overlayfs and union-mounts behave as if copy-up was a > "bind mount" over the file in question. Yes, it's a namespace trick > but it seems to work in most situations. It must be a sad news for overlayfs users who have every met "non-most" situations. > The point was that uid/gid was to be overridden with *different* > values for each overlayfs instance. So implementing uid/gid > overriding in the lower fs doesn't help. For the scenario which Alessandro Pignotti posted, - everything on the lower fs are owned by a single user, by the overriding feature. - everything on the upper fs are owned by the user natively. Isn't it enough? J. R. Okajima