[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4ECD6375.5050404@suse.cz>
Date: Wed, 23 Nov 2011 22:19:49 +0100
From: Jiri Slaby <jslaby@...e.cz>
To: Havard Skinnemoen <hskinnemoen@...gle.com>
CC: Jiri Slaby <jirislaby@...il.com>, Dave Jones <davej@...hat.com>,
Linux Kernel <linux-kernel@...r.kernel.org>,
Greg Kroah-Hartman <gregkh@...e.de>,
Alan Cox <alan@...ux.intel.com>
Subject: Re: [RFC] cdc-acm: Fix potential deadlock (lockdep warning)
On 11/23/2011 10:11 PM, Jiri Slaby wrote:
> On 11/23/2011 10:08 PM, Havard Skinnemoen wrote:
>> tty_port_close_start() takes port->lock, so if we try to do that,
>> we'll turn a potential deadlock into a real one. Or did you mean
>> something else?
>
> I mean the other uses of port->count in your driver.
>
>> I think I'll try to follow Alan's advice and use tty_port_close()
>> instead of the split functions.
>
> Yes, that's indeed the best approach. However you have to use them all:
> tty_port_open, tty_port_hangup, tty_port_close. OR you would have to do
> the locking of count properly.
And I see now that the driver uses tty_port_hangup (count under
port->lock) and tty_port_close_start (count under open_mutex and
port->lock). But doesn't use tty_port_open (count under open_mutex and
acm->mutex). Hence there was always a race hangup vs. open wrt. port
counting.
thanks,
--
js
suse labs
--
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/
Powered by blists - more mailing lists