[PATCH V3 2/2] staging/comedi/drivers: free allocated priv and release resources if comedi_alloc_subdevices

[PATCH V3 2/2] staging/comedi/drivers: free allocated priv and release resources if comedi_alloc_subdevices

[Devendra Naga]

as comedi_alloc_subdevices may fail, we leak out the memory allocated by alloc_private,
and also the I/O regions we requested.

free out the private data pointer of the driver, and release the I/O regions

Signed-off-by: Devendra Naga <devendra.aaru@gmail.com>
---
 drivers/staging/comedi/drivers/fl512.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/staging/comedi/drivers/fl512.c b/drivers/staging/comedi/drivers/fl512.c
index 52e6d14..1c372d1 100644
--- a/drivers/staging/comedi/drivers/fl512.c
+++ b/drivers/staging/comedi/drivers/fl512.c
@@ -135,8 +135,11 @@ static int fl512_attach(struct comedi_device *dev, struct comedi_devconfig *it)
 #endif
 
 	ret = comedi_alloc_subdevices(dev, 2);
-	if (ret)
+	if (ret) {
+		kfree(dev->private);
+		release_region(iobase, FL512_SIZE);
 		return ret;
+	}
 
 	/*
 	 * this if the definitions of the supdevices, 2 have been defined
-- 
1.7.9.5

Re: [PATCH V3 2/2] staging/comedi/drivers: free allocated priv and release resources if comedi_alloc_subdevices

[Ian Abbott]

On 2012-07-01 17:20, Devendra Naga wrote:
> as comedi_alloc_subdevices may fail, we leak out the memory allocated by alloc_private,
> and also the I/O regions we requested.
>
> free out the private data pointer of the driver, and release the I/O regions
>
> Signed-off-by: Devendra Naga <devendra.aaru@gmail.com>
> ---
>   drivers/staging/comedi/drivers/fl512.c |    5 ++++-
>   1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/staging/comedi/drivers/fl512.c b/drivers/staging/comedi/drivers/fl512.c
> index 52e6d14..1c372d1 100644
> --- a/drivers/staging/comedi/drivers/fl512.c
> +++ b/drivers/staging/comedi/drivers/fl512.c
> @@ -135,8 +135,11 @@ static int fl512_attach(struct comedi_device *dev, struct comedi_devconfig *it)
>   #endif
>
>   	ret = comedi_alloc_subdevices(dev, 2);
> -	if (ret)
> +	if (ret) {
> +		kfree(dev->private);
> +		release_region(iobase, FL512_SIZE);
>   		return ret;
> +	}
>
>   	/*
>   	 * this if the definitions of the supdevices, 2 have been defined
>

As mentioned in my comment for your PATCH 1/2, this is unneccessary.  In 
fact it will resuly in dev->private being freed twice and the I/O region 
being released twice due to clean-up in the comedi core when 
fl512_attach() returns an error, which will also call fl512_detach() 
during the clean-up.

-- 
-=( Ian Abbott @ MEV Ltd.    E-mail: <abbotti@mev.co.uk>        )=-
-=( Tel: +44 (0)161 477 1898   FAX: +44 (0)161 718 3587         )=-

Re: [PATCH V3 2/2] staging/comedi/drivers: free allocated priv and release resources if comedi_alloc_subdevices

[devendra.aaru]

On Mon, Jul 2, 2012 at 2:15 PM, Ian Abbott <abbotti@mev.co.uk> wrote:
> On 2012-07-01 17:20, Devendra Naga wrote:
>>
>> as comedi_alloc_subdevices may fail, we leak out the memory allocated by
>> alloc_private,
>> and also the I/O regions we requested.
>>
>> free out the private data pointer of the driver, and release the I/O
>> regions
>>
>> Signed-off-by: Devendra Naga <devendra.aaru@gmail.com>
>> ---
>>   drivers/staging/comedi/drivers/fl512.c |    5 ++++-
>>   1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/staging/comedi/drivers/fl512.c
>> b/drivers/staging/comedi/drivers/fl512.c
>> index 52e6d14..1c372d1 100644
>> --- a/drivers/staging/comedi/drivers/fl512.c
>> +++ b/drivers/staging/comedi/drivers/fl512.c
>> @@ -135,8 +135,11 @@ static int fl512_attach(struct comedi_device *dev,
>> struct comedi_devconfig *it)
>>   #endif
>>
>>         ret = comedi_alloc_subdevices(dev, 2);
>> -       if (ret)
>> +       if (ret) {
>> +               kfree(dev->private);
>> +               release_region(iobase, FL512_SIZE);
>>                 return ret;
>> +       }
>>
>>         /*
>>          * this if the definitions of the supdevices, 2 have been defined
>>
>
> As mentioned in my comment for your PATCH 1/2, this is unneccessary.  In
> fact it will resuly in dev->private being freed twice and the I/O region
> being released twice due to clean-up in the comedi core when fl512_attach()
> returns an error, which will also call fl512_detach() during the clean-up.
>
> --
> -=( Ian Abbott @ MEV Ltd.    E-mail: <abbotti@mev.co.uk>        )=-
> -=( Tel: +44 (0)161 477 1898   FAX: +44 (0)161 718 3587         )=-
>
>

Ok, sorry for the heavy noise caused by these patches :(.

anyway i have a question

are you going to remove the alloc_private wapper and use kzalloc?

Thanks,
Devendra.

Re: [PATCH V3 2/2] staging/comedi/drivers: free allocated priv and release resources if comedi_alloc_subdevices

[devendra.aaru]

On Mon, Jul 2, 2012 at 2:48 PM, Ian Abbott <abbotti@mev.co.uk> wrote:
> On 2012-07-02 09:51, devendra.aaru wrote:
>>
>> anyway i have a question
>>
>> are you going to remove the alloc_private wapper and use kzalloc?
>
>
> I'm not averse to the idea, but it's not a high priority.
>

Then will i do? and change all drivers?
I dunno about any other guys doing it ...
>
> --
> -=( Ian Abbott @ MEV Ltd.    E-mail: <abbotti@mev.co.uk>        )=-
> -=( Tel: +44 (0)161 477 1898   FAX: +44 (0)161 718 3587         )=-
>
>

Thanks,
Devendra.