linux-kernel - Re: [PATCH AUTOSEL 4.19 12/32] Btrfs: fix deadlock between fiemap and transaction commits

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20190807124437.GT28208@twin.jikos.cz>
Date:   Wed, 7 Aug 2019 14:44:37 +0200
From:   David Sterba <dsterba@...e.cz>
To:     Greg KH <gregkh@...uxfoundation.org>
Cc:     dsterba@...e.cz, Sasha Levin <sashal@...nel.org>,
        linux-kernel@...r.kernel.org, stable@...r.kernel.org,
        David Sterba <dsterba@...e.com>,
        Filipe Manana <fdmanana@...e.com>, linux-btrfs@...r.kernel.org
Subject: Re: [PATCH AUTOSEL 4.19 12/32] Btrfs: fix deadlock between fiemap
 and transaction commits

On Wed, Aug 07, 2019 at 12:51:26PM +0200, Greg KH wrote:
> On Wed, Aug 07, 2019 at 11:47:59AM +0200, David Sterba wrote:
> > On Tue, Aug 06, 2019 at 05:35:00PM -0400, Sasha Levin wrote:
> > > From: Filipe Manana <fdmanana@...e.com>
> > > 
> > > [ Upstream commit a6d155d2e363f26290ffd50591169cb96c2a609e ]
> > > 
> > > Fixes: 03628cdbc64db6 ("Btrfs: do not start a transaction during fiemap")
> > 
> > The commit is a regression fix during the 5.2 cycle, how it could end up
> > in a 4.19 stable candidate?
> > 
> > $ git describe  03628cdbc64db6
> > v5.1-rc7-201-g03628cdbc64d
> > 
> > $ git describe --contains 03628cdbc64db6
> > v5.2-rc1~163^2~26
> > 
> > And it does not belong to 5.2 either, git cherry-pick on top of 5.2
> > fails.
> > 
> > I think such sanity check can be done automatically so the patches don't
> > accidentally land in trees where don't belong.
> 
> 
> Commit 03628cdbc64d ("Btrfs: do not start a transaction during fiemap")
> was tagged for the stable trees, and ended up in the following releases:
> 	4.14.121 4.19.45 5.0.18 5.1.4 5.2
> so yes, it does need to go back to all of those locations if this patch
> really does fix the issue there.

You're right, I did not notice the CC tag when examining the patches.