[RFC PATCH v1 4/8] zram: destroy all devices on error recovery path in zram_init()

Jiang Liu liuj97 at gmail.com
Tue Jun 4 14:57:23 UTC 2013


On Tue 04 Jun 2013 04:49:16 PM CST, Dan Carpenter wrote:
> Everyone stop putting RFC on their bugfixes!  :P  No one wants to
> pre-review patches.
>
> On Mon, Jun 03, 2013 at 11:42:16PM +0800, Jiang Liu wrote:
>> On error recovery path of zram_init(), it leaks the zram device object
>> causing the failure.
>>
>
> This is a real bug but the fix isn't right.  The object causing the
> failure has only been partially set up.  And in some cases it has
> been partially cleaned up, for example we could end up releasing
> ->queue twice.
>
> The better way is to make create_device() clean up after itself
> completely instead of only partly and sometimes.
>
> regards,
> dan carpenter
>
Hi Dan,
     Thanks for your review, will fix it in next version.
Regards!
Gerry




More information about the devel mailing list