linux-kernel - Re: JFFS2 deadlock

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-next>] [day] [month] [year] [list]

Message-ID: <OF19585295.84AA7860-ONC1257F5D.0031E02A-C1257F5D.0036B10E@rohde-schwarz.com>
Date:	Thu, 18 Feb 2016 10:57:21 +0100
From:	Thomas.Betker@...de-schwarz.com
To:	David Woodhouse <dwmw2@...radead.org>
Cc:	computersforpeace@...il.com,
	Artem Bityutskiy <dedekind1@...il.com>,
	Deng Chao <deng.chao1@....com.cn>,
	Joakim Tjernlund <Joakim.Tjernlund@...inera.com>,
	linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
	"linux-mtd@...ts.infradead.org" <linux-mtd@...ts.infradead.org>,
	linux-mtd <linux-mtd-bounces@...ts.infradead.org>,
	Ming Liu <liu.ming50@...il.com>,
	"sztomi89@...il.com" <sztomi89@...il.com>,
	Alexander Viro <viro@...iv.linux.org.uk>,
	wangzaiwei <wangzaiwei@...-vision.cn>
Subject: Re: JFFS2 deadlock

Hello David:

> > Please could you try what's in the tree at
> > http://git.infradead.org/users/dwmw2/jffs2-fixes.git

> Your patch looks much simpler, and I will definitely test it. It may
> take a few days, though, as I have to unearth the test scripts, and 
> find a time slot for testing.

Here is what I did (sorry for the wait, things were piling up):

1) Removed Deng Chao's patch from my kernel, added your patch "jffs2: Fix 
page lock / f->sem deadlock". I am still on linux-3.14, but jffs2 hasn't 
changed much since then, so this shouldn't make a difference. Added a 
printk() before mutex_unlock(&f->sem) to check if the prospective page was 
locked, i.e. if the deadlock situation actually occurs.

2) On my target system, started wangzaiwei's test (with some fixes), plus 
a loop copying a large file over and over (to get GC rolling, which 
increases the chance of a deadlock).

3) After 24 hours, the system was still alive, and the printk() had been 
hit 32 times.

So yes, I am confident that your patch avoids the deadlock, and if that's 
good enough for you, please add my Tested-by:. However, I am going to run 
some more stress tests here just to check that there weren't any 
unexpected side effects. (Don't get me wrong -- I am sure the patch is 
fine, but for me it's a case of "once bitten, twice shy" ...)

Best regards,
Thomas Betker