linux-kernel - Re: Oops after merge of tty-next

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: <20120727071634.GA29885@avionic-0098.mockup.avionic-design.de>
Date:	Fri, 27 Jul 2012 09:16:34 +0200
From:	Thierry Reding <thierry.reding@...onic-design.de>
To:	Alan Cox <alan@...ux.intel.com>
Cc:	Ian Abbott <abbotti@....co.uk>,
	Alan Cox <alan@...rguk.ukuu.org.uk>,
	Ian Abbott <ian.abbott@....co.uk>,
	lkml <linux-kernel@...r.kernel.org>,
	Greg KH <gregkh@...uxfoundation.org>
Subject: Re: Oops after merge of tty-next

On Tue, Jul 24, 2012 at 11:05:27PM +0100, Alan Cox wrote:
> On Mon, 23 Jul 2012 15:51:03 +0100
> Ian Abbott <abbotti@....co.uk> wrote:
> 
> > On 2012-07-21 23:41, Alan Cox wrote:
> > > On Fri, 20 Jul 2012 23:07:06 +0100
> > > Ian Abbott <abbotti@....co.uk> wrote:
> > >
> > >> I'm getting an Oops in the linux-next tree today after the merge
> > >> of the remote-tracking branch 'tty/tty-next'.  I bisected it down
> > >> to commit 36b3c070d2346c890d690d71f6eab02f8c511137 in
> > >> git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git :
> 
> Ok Greg you can you leave that one in -next but not push it for 3.6.
> 
> I need to go over this in some detail and figure out the remaining
> race, and worse yet how to fix it without the mess of existing locks
> turning it into something nasty.
> 
> I think I understand what is needed but I don't want to be doing it as
> a mad panic for 3.6. On the bright side I think it explains the other
> tty lock splitting mysteries.

I've also been able to reproduce this, or at least a very similar issue,
on ARM (Tegra). The system boots into the initrd, which works as usual,
but it hangs when executing the switch_root that starts systemd within
the final root filesystem. With the above-mentioned commit reverted, the
system successfully boots to the login prompt.

Thierry

Content of type "application/pgp-signature" skipped