2022-01-06 08:31:31

by Xianting Tian

Subject: [PATCH] virtio-pci: fix memory leak of vp_dev

In virtio_pci_probe(), vp_dev will not be freed if it goes to
"err_register" branch. fix it

Signed-off-by: Xianting Tian <[email protected]>
---
drivers/virtio/virtio_pci_common.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/virtio/virtio_pci_common.c b/drivers/virtio/virtio_pci_common.c
index fdbde1db5..48b5ac15e 100644
--- a/drivers/virtio/virtio_pci_common.c
+++ b/drivers/virtio/virtio_pci_common.c
@@ -602,8 +602,7 @@ static int virtio_pci_probe(struct pci_dev *pci_dev,
err_enable_device:
if (reg_dev)
put_device(&vp_dev->vdev.dev);
- else
- kfree(vp_dev);
+ kfree(vp_dev);
return rc;
}

--
2.17.1

2022-01-07 02:34:22

by Jason Wang

Subject: Re: [PATCH] virtio-pci: fix memory leak of vp_dev

On Thu, Jan 6, 2022 at 4:31 PM Xianting Tian
<[email protected]> wrote:
>
> In virtio_pci_probe(), vp_dev will not be freed if it goes to
> "err_register" branch. fix it
>
> Signed-off-by: Xianting Tian <[email protected]>

So we had this:

commit 33635bd976fb4c3ccf0cfbb81a8d29bb87760607
Author: weiping zhang <[email protected]>
Date: Thu Dec 21 20:40:24 2017 +0800

virtio_pci: don't kfree device on register failure

As mentioned at drivers/base/core.c:
/*
* NOTE: _Never_ directly free @dev after calling this function, even
* if it returned an error! Always use put_device() to give up the
* reference initialized in this function instead.
*/
so we don't free vp_dev until vp_dev->vdev.dev.release be called.

Signed-off-by: weiping zhang <[email protected]>
Reviewed-by: Cornelia Huck <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

Any reason that the above fix is wrong?

Thanks

> ---
> drivers/virtio/virtio_pci_common.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/drivers/virtio/virtio_pci_common.c b/drivers/virtio/virtio_pci_common.c
> index fdbde1db5..48b5ac15e 100644
> --- a/drivers/virtio/virtio_pci_common.c
> +++ b/drivers/virtio/virtio_pci_common.c
> @@ -602,8 +602,7 @@ static int virtio_pci_probe(struct pci_dev *pci_dev,
> err_enable_device:
> if (reg_dev)
> put_device(&vp_dev->vdev.dev);
> - else
> - kfree(vp_dev);
> + kfree(vp_dev);
> return rc;
> }
>
> --
> 2.17.1
>

2022-01-07 02:36:53

by Xianting Tian

Subject: Re: [PATCH] virtio-pci: fix memory leak of vp_dev

Thanks, I checked this before your reply, it is not a bug,

Sorry for the inconvenience caused.

在 2022/1/7 上午10:34, Jason Wang 写道:
> On Thu, Jan 6, 2022 at 4:31 PM Xianting Tian
> <[email protected]> wrote:
>> In virtio_pci_probe(), vp_dev will not be freed if it goes to
>> "err_register" branch. fix it
>>
>> Signed-off-by: Xianting Tian <[email protected]>
> So we had this:
>
> commit 33635bd976fb4c3ccf0cfbb81a8d29bb87760607
> Author: weiping zhang <[email protected]>
> Date: Thu Dec 21 20:40:24 2017 +0800
>
> virtio_pci: don't kfree device on register failure
>
> As mentioned at drivers/base/core.c:
> /*
> * NOTE: _Never_ directly free @dev after calling this function, even
> * if it returned an error! Always use put_device() to give up the
> * reference initialized in this function instead.
> */
> so we don't free vp_dev until vp_dev->vdev.dev.release be called.
>
> Signed-off-by: weiping zhang <[email protected]>
> Reviewed-by: Cornelia Huck <[email protected]>
> Signed-off-by: Michael S. Tsirkin <[email protected]>
>
> Any reason that the above fix is wrong?
>
> Thanks
>
>> ---
>> drivers/virtio/virtio_pci_common.c | 3 +--
>> 1 file changed, 1 insertion(+), 2 deletions(-)
>>
>> diff --git a/drivers/virtio/virtio_pci_common.c b/drivers/virtio/virtio_pci_common.c
>> index fdbde1db5..48b5ac15e 100644
>> --- a/drivers/virtio/virtio_pci_common.c
>> +++ b/drivers/virtio/virtio_pci_common.c
>> @@ -602,8 +602,7 @@ static int virtio_pci_probe(struct pci_dev *pci_dev,
>> err_enable_device:
>> if (reg_dev)
>> put_device(&vp_dev->vdev.dev);
>> - else
>> - kfree(vp_dev);
>> + kfree(vp_dev);
>> return rc;
>> }
>>
>> --
>> 2.17.1
>>