Hi all,

I found that when we release eq->kobj by invoking kobject_put() we had gave
misson to workqueue, so that workqueue will release the elevator queue
space in elevator_alloc() which is registered in ->release of elv_ktype object.
We should not release it again.

So, this fix is wrong, please ignore this patch, sorry for the noisy!

Regards,

> -----Original Message-----
> From: Chao Yu [mailto:chao2...@samsung.com]
> Sent: Friday, April 24, 2015 10:09 AM
> To: 'ax...@kernel.dk'
> Cc: 'linux-kernel@vger.kernel.org'
> Subject: [PATCH] elevator: fix memory leak in ->elevator_init_fn
> 
> In ->elevator_init_fn, if we fail to call kzalloc_node, we should release
> elevator queue space which is allocated previously, otherwise it will cause
> memory leak.
> 
> Signed-off-by: Chao Yu <chao2...@samsung.com>
> ---
>  block/cfq-iosched.c      | 1 +
>  block/deadline-iosched.c | 1 +
>  block/noop-iosched.c     | 1 +
>  3 files changed, 3 insertions(+)
> 
> diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
> index 5da8e6e..2793fb7 100644
> --- a/block/cfq-iosched.c
> +++ b/block/cfq-iosched.c
> @@ -4391,6 +4391,7 @@ static int cfq_init_queue(struct request_queue *q, 
> struct elevator_type
> *e)
>       cfqd = kzalloc_node(sizeof(*cfqd), GFP_KERNEL, q->node);
>       if (!cfqd) {
>               kobject_put(&eq->kobj);
> +             kfree(eq);
>               return -ENOMEM;
>       }
>       eq->elevator_data = cfqd;
> diff --git a/block/deadline-iosched.c b/block/deadline-iosched.c
> index a753df2..bc6d7d9 100644
> --- a/block/deadline-iosched.c
> +++ b/block/deadline-iosched.c
> @@ -349,6 +349,7 @@ static int deadline_init_queue(struct request_queue *q, 
> struct elevator_type
> *e)
>       dd = kzalloc_node(sizeof(*dd), GFP_KERNEL, q->node);
>       if (!dd) {
>               kobject_put(&eq->kobj);
> +             kfree(eq);
>               return -ENOMEM;
>       }
>       eq->elevator_data = dd;
> diff --git a/block/noop-iosched.c b/block/noop-iosched.c
> index 3de89d4..1399c78 100644
> --- a/block/noop-iosched.c
> +++ b/block/noop-iosched.c
> @@ -71,6 +71,7 @@ static int noop_init_queue(struct request_queue *q, struct 
> elevator_type *e)
>       nd = kmalloc_node(sizeof(*nd), GFP_KERNEL, q->node);
>       if (!nd) {
>               kobject_put(&eq->kobj);
> +             kfree(eq);
>               return -ENOMEM;
>       }
>       eq->elevator_data = nd;
> --
> 2.3.3


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Reply via email to