Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH] ssb:: use put_device() if device_register fail
To:     =?UTF-8?Q?Michael_B=c3=bcsch?= <m@bues.ch>
References: <bdd20c00cba68044d82e387072b39f0176448e97.1520416785.git.arvind.yadav.cs@gmail.com>
 <20180307174700.013e20ae@wiggum> <5AA01E5E.8070002@gmail.com>
 <20180307183858.63172de9@wiggum>
Cc:     linux-kernel@vger.kernel.org, linux-wireless@vger.kernel.org
From:   arvindY <arvind.yadav.cs@gmail.com>
Message-ID: <5AA02DD2.2090709@gmail.com>
Date:   Wed, 7 Mar 2018 23:52:10 +0530
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101
 Thunderbird/38.6.0
MIME-Version: 1.0
In-Reply-To: <20180307183858.63172de9@wiggum>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


On Wednesday 07 March 2018 11:08 PM, Michael B?sch wrote:
> On Wed, 7 Mar 2018 22:46:14 +0530
> arvindY <arvind.yadav.cs@gmail.com> wrote:
>>>> diff --git a/drivers/ssb/main.c b/drivers/ssb/main.c
>>>> index 65420a9..c4449e0 100644
>>>> --- a/drivers/ssb/main.c
>>>> +++ b/drivers/ssb/main.c
>>>> @@ -521,6 +521,7 @@ static int ssb_devices_register(struct ssb_bus *bus)
>>>>    			ssb_err("Could not register %s\n", dev_name(dev));
>>>>    			/* Set dev to NULL to not unregister
>>>>    			 * dev on error unwinding. */
>>>> +			put_device(dev);
>>>>    			sdev->dev = NULL;
>>>>    			kfree(devwrap);
>>>>    			goto error;
>>> I don't think this is correct.
>>> The dev structure is allocated as part of devwrap, which is freed here.
>>>
>>> Why do you think we need put_device here?
>>>   
>> Yes this patch is not correct, We must not use kfree() after you called
>> device_register() (even
>> if it was not successful!) -- see the comment for device_register().
>> I will delete kfree() and send updated patch.
>
> Is device_put() going to call ssb_release_dev() to free the structure?
>
> Can you please elaborate on why device_put() must be used? The comment
> is not really of any use here.
>
put_device() will call kobject_put(). By calling this, The kobject core 
will automatically
clean up all of the memory allocated with the kobject. Internally 
kobject_put() will call
kobject_cleanup() which is responsible to call 'dev -> release' and also 
free other
kobject resources. we should always avoid kfree() if device_register() 
returned an error.
Otherwise it'll not do clean up of other kobject resources.

~arvind