From mboxrd@z Thu Jan 1 00:00:00 1970 From: Shmulik Ladkani Subject: Re: [PATCH 1/2] Revert "af_unix: Fix splice-bind deadlock" Date: Sat, 3 Sep 2016 20:00:13 +0300 Message-ID: <20160903200013.14111d91@halley> References: Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: "David S. Miller" , Hannes Frederic Sowa , Rainer Weikusat , Eric Dumazet , willy tarreau , netdev@vger.kernel.org To: Linus Torvalds Return-path: Received: from mail-wm0-f68.google.com ([74.125.82.68]:36420 "EHLO mail-wm0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753422AbcICRA1 (ORCPT ); Sat, 3 Sep 2016 13:00:27 -0400 Received: by mail-wm0-f68.google.com with SMTP id i138so6989723wmf.3 for ; Sat, 03 Sep 2016 10:00:26 -0700 (PDT) In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Fri, 2 Sep 2016 11:09:23 -0700 (PDT) Linus Torvalds wrote: > From: Linus Torvalds > Date: Thu, 1 Sep 2016 14:56:49 -0700 > Subject: [PATCH 1/2] Revert "af_unix: Fix splice-bind deadlock" > > This reverts commit c845acb324aa85a39650a14e7696982ceea75dc1. > > It turns out that it just replaces one deadlock with another one: we can > still get the wrong lock ordering with the readlock due to overlayfs > calling back into the filesystem layer and still taking the vfs locks > after the readlock. > > The proper solution ends up being to just split the readlock into two > pieces: the bind lock (taken *outside* the vfs locks) and the IO lock > (taken *inside* the filesystem locks). The two locks are independent > anyway. > > Signed-off-by: Linus Torvalds > --- > > This is not a completely clean revert, because other changes had happened > in this area since that commit, but the conflicts were pretty trivial. Reviewed-by: Shmulik Ladkani