From mboxrd@z Thu Jan  1 00:00:00 1970
From: Al Viro <viro@ZenIV.linux.org.uk>
Subject: Re: Intentionally corrupted vfat fs causing BUG
Date: Thu, 23 Oct 2014 19:05:11 +0100
Message-ID: <20141023180510.GG7996@ZenIV.linux.org.uk>
References: <87y4sk5pul.fsf@devron.myhome.or.jp>
 <543B8FA7.9000106@nod.at>
 <87r3yc5oqt.fsf@devron.myhome.or.jp>
 <5443E87A.2060207@nod.at>
 <87oat29551.fsf@devron.myhome.or.jp>
 <20141023160106.GB7996@ZenIV.linux.org.uk>
 <20141023161606.GC7996@ZenIV.linux.org.uk>
 <877fzq91ky.fsf@devron.myhome.or.jp>
 <20141023165533.GD7996@ZenIV.linux.org.uk>
 <20141023175428.GB22339@fieldses.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>,
	Richard Weinberger <richard@nod.at>,
	Sami Liedes <sami.liedes@iki.fi>,
	linux-fsdevel <linux-fsdevel@vger.kernel.org>
To: "J. Bruce Fields" <bfields@fieldses.org>
Return-path: <linux-fsdevel-owner@vger.kernel.org>
Received: from zeniv.linux.org.uk ([195.92.253.2]:37622 "EHLO
	ZenIV.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S932090AbaJWSFP (ORCPT
	<rfc822;linux-fsdevel@vger.kernel.org>);
	Thu, 23 Oct 2014 14:05:15 -0400
Content-Disposition: inline
In-Reply-To: <20141023175428.GB22339@fieldses.org>
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Thu, Oct 23, 2014 at 01:54:28PM -0400, J. Bruce Fields wrote:

> > Hmm...  Not in the current mainline (and not because of want_discon - that's
> > gone already).  However, with the fixes I've got in the local tree it
> > will both find and move it - same as d_materialise_unique() would in the
> > current mainline.
> 
> What happened to the idea of failing with -EIO in that case?

Quiet d_move() is definitely better in this case.  Failing with EIO is what
d_splice_alias() does *now*, and that's why the patch I've posted uses
d_materialise_unique().  Patches in my local tree switch that case of
d_splice_alias() to what d_materialise_unique() does (and kill
d_materialise_unique() completely - compat #define is left, but all
in-tree users are switched to d_splice_alias())...