linux-kernel - Re: w1: 3.14-rc7 - possible recursive locking detected

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

Date:	Mon, 14 Apr 2014 08:04:45 -0700
From:	Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:	David Fries <david@...es.net>
Cc:	Belisko Marek <marek.belisko@...il.com>,
	Evgeniy Polyakov <zbr@...emap.net>,
	LKML <linux-kernel@...r.kernel.org>,
	"Dr. H. Nikolaus Schaller" <hns@...delico.com>
Subject: Re: w1: 3.14-rc7 - possible recursive locking detected

On Sun, Apr 13, 2014 at 05:24:01PM -0500, David Fries wrote:
> Belisko Marek,
> Here is a possible solution, could you give it a try and report back?
> 
> Greg Kroah-Hartman,
> Evgeniy asked me to look into this report.  I don't have the
> reporter's hardware configuration, but I wouldn't think that would be
> needed, just some w1 bus master (even W1_MASTER_GPIO might work), then
> loading the slave device and manually adding a slave device with that
> family id.  Even then I didn't reproduce the reported recursive
> locking error.  I saw unrelated locking reports, but not this one.  I
> wrote up the included patch, but that undoes the notify changes that
> you added earlier in commit 47eba33a0997fc7, and I wanted to ask about
> that commit.  Specifically these two lines,
> 
>         err = device_register(&sl->dev);
> 	...
> +	dev_set_uevent_suppress(&sl->dev, false);
> +	kobject_uevent(&sl->dev.kobj, KOBJ_ADD);
> 
> Wouldn't the default be to not suppress?  Nothing in the W1 system
> enables suppressing so is that even needed?  (I'm fine with saying
> it's a good idea).
> device_register at some point must call device_add and device_add
> calls kobject_uevent(&dev->kobj, KOBJ_ADD);
> so doesn't the KOBJ_ADD send the add a second time?  As in it
> shouldn't be needed?
> Can the suppress be called before device_register to avoid the
> automatic notify, then after it returns setup the slave device as this
> patch does to avoid this problem report, and then call the KOBJ_ADD to
> make everything happy?

I really have no idea, if your fix resolves an issue, that's great, I'll
be glad to take it.  I have no w1 devices to test any of this, and don't
even remember writing that kernel patch :)

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/