From mboxrd@z Thu Jan  1 00:00:00 1970
From: andre.przywara@arm.com (Andre Przywara)
Date: Wed, 08 Apr 2015 16:39:03 +0100
Subject: [PATCH v2 01/10] drivers: PL011: avoid potential unregister_driver
 call
In-Reply-To: <20150312104258.GI8656@n2100.arm.linux.org.uk>
References: <1425491994-23913-1-git-send-email-andre.przywara@arm.com>
 <1425491994-23913-2-git-send-email-andre.przywara@arm.com>
 <20150312104258.GI8656@n2100.arm.linux.org.uk>
Message-ID: <55254B97.6050704@arm.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Hi Russell,

On 12/03/15 10:42, Russell King - ARM Linux wrote:
> On Wed, Mar 04, 2015 at 05:59:45PM +0000, Andre Przywara wrote:
>> Although we care about not unregistering the driver if there are
>> still ports connected during the .remove callback, we do miss this
>> check in the pl011_probe function. So if the current port allocation
>> fails, but there are other ports already registered, we will kill
>> those.
>> So factor out the port removal into a separate function and use that
>> in the probe function, too.
>>
>> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
>> ---
>>  drivers/tty/serial/amba-pl011.c |   38 +++++++++++++++++++++-----------------
>>  1 file changed, 21 insertions(+), 17 deletions(-)
>>
>> diff --git a/drivers/tty/serial/amba-pl011.c b/drivers/tty/serial/amba-pl011.c
>> index 92783fc..961f9b0 100644
>> --- a/drivers/tty/serial/amba-pl011.c
>> +++ b/drivers/tty/serial/amba-pl011.c
>> @@ -2235,6 +2235,24 @@ static int pl011_probe_dt_alias(int index, struct device *dev)
>>  	return ret;
>>  }
>>  
>> +/* unregisters the driver also if no more ports are left */
>> +static void pl011_unregister_port(struct uart_amba_port *uap)
>> +{
>> +	int i;
>> +	bool busy = false;
>> +
>> +	for (i = 0; i < ARRAY_SIZE(amba_ports); i++) {
>> +		if (amba_ports[i] == uap)
>> +			amba_ports[i] = NULL;
>> +		else if (amba_ports[i])
>> +			busy = true;
>> +	}
>> +	pl011_dma_remove(uap);
>> +	if (!busy)
>> +		uart_unregister_driver(&amba_reg);
>> +}
> 
> This is still racy, as I pointed out at the time this crap was dreamt
> up.
> 
> There is _no_ locking between an individual driver's ->probe or ->remove
> functions being called concurrently for different devices.  The only
> locking which the driver model guarantees is that a single struct device
> can only be probed by one driver at a time.
> 
> Multiple struct device's can be in-progress of ->probe or ->remove
> simultaneously.

OK, I see.

> However, this isn't your bug to solve... it's those who were proponents
> of this crap approach.

Does that mean you want me to drop this patch? It isn't strictly
necessary for my series. So do you want to postpone a fix until later
when there is a real solution (tm) for this issue or shall I include
this still in my series for fixing at least half of the issue?

Thanks,
Andre.