2024-02-06 07:38:33

by Zhipeng Lu

Subject: [PATCH] media: ttpci: fix two memleaks in budget_av_attach

When saa7146_register_device and saa7146_vv_init fails, budget_av_attach
should free the resources it allocates, like the error-handling of
ttpci_budget_init does. Besides, there are two fixme comment refers to
such deallocations.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Zhipeng Lu <[email protected]>
---
drivers/media/pci/ttpci/budget-av.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/media/pci/ttpci/budget-av.c b/drivers/media/pci/ttpci/budget-av.c
index 230b104a7cdf..4f897f848581 100644
--- a/drivers/media/pci/ttpci/budget-av.c
+++ b/drivers/media/pci/ttpci/budget-av.c
@@ -1463,7 +1463,8 @@ static int budget_av_attach(struct saa7146_dev *dev, struct saa7146_pci_extensio
budget_av->has_saa7113 = 1;
err = saa7146_vv_init(dev, &vv_data);
if (err != 0) {
- /* fixme: proper cleanup here */
+ ttpci_budget_deinit(&budget_av->budget);
+ kfree(budget_av);
ERR("cannot init vv subsystem\n");
return err;
}
@@ -1472,7 +1473,8 @@ static int budget_av_attach(struct saa7146_dev *dev, struct saa7146_pci_extensio
vv_data.vid_ops.vidioc_s_input = vidioc_s_input;

if ((err = saa7146_register_device(&budget_av->vd, dev, "knc1", VFL_TYPE_VIDEO))) {
- /* fixme: proper cleanup here */
+ ttpci_budget_deinit(&budget_av->budget);
+ kfree(budget_av);
ERR("cannot register capture v4l2 device\n");
saa7146_vv_release(dev);
return err;
--
2.34.1

2024-02-14 09:28:06

by Hans Verkuil

Subject: Re: [PATCH] media: ttpci: fix two memleaks in budget_av_attach

On 06/02/2024 08:37, Zhipeng Lu wrote:
> When saa7146_register_device and saa7146_vv_init fails, budget_av_attach
> should free the resources it allocates, like the error-handling of
> ttpci_budget_init does. Besides, there are two fixme comment refers to
> such deallocations.
>
> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> Signed-off-by: Zhipeng Lu <[email protected]>
> ---
> drivers/media/pci/ttpci/budget-av.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/media/pci/ttpci/budget-av.c b/drivers/media/pci/ttpci/budget-av.c
> index 230b104a7cdf..4f897f848581 100644
> --- a/drivers/media/pci/ttpci/budget-av.c
> +++ b/drivers/media/pci/ttpci/budget-av.c
> @@ -1463,7 +1463,8 @@ static int budget_av_attach(struct saa7146_dev *dev, struct saa7146_pci_extensio
> budget_av->has_saa7113 = 1;
> err = saa7146_vv_init(dev, &vv_data);
> if (err != 0) {
> - /* fixme: proper cleanup here */
> + ttpci_budget_deinit(&budget_av->budget);
> + kfree(budget_av);
> ERR("cannot init vv subsystem\n");
> return err;
> }
> @@ -1472,7 +1473,8 @@ static int budget_av_attach(struct saa7146_dev *dev, struct saa7146_pci_extensio
> vv_data.vid_ops.vidioc_s_input = vidioc_s_input;
>
> if ((err = saa7146_register_device(&budget_av->vd, dev, "knc1", VFL_TYPE_VIDEO))) {
> - /* fixme: proper cleanup here */
> + ttpci_budget_deinit(&budget_av->budget);
> + kfree(budget_av);
I think this should go after the saa7146_vv_release() release below.
Just in case saa7146_vv_release() relies on budget_av somewhere. And
besides, it is good practice to release resources in the reverse order
they were allocated.

Regards,

Hans

> ERR("cannot register capture v4l2 device\n");
> saa7146_vv_release(dev);
> return err;

2024-02-21 05:07:42

by Zhipeng Lu

Subject: Re: [PATCH] media: ttpci: fix two memleaks in budget_av_attach

> On 06/02/2024 08:37, Zhipeng Lu wrote:
> > When saa7146_register_device and saa7146_vv_init fails, budget_av_attach
> > should free the resources it allocates, like the error-handling of
> > ttpci_budget_init does. Besides, there are two fixme comment refers to
> > such deallocations.
> >
> > Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
> > Signed-off-by: Zhipeng Lu <[email protected]>
> > ---
> > drivers/media/pci/ttpci/budget-av.c | 6 ++++--
> > 1 file changed, 4 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/media/pci/ttpci/budget-av.c b/drivers/media/pci/ttpci/budget-av.c
> > index 230b104a7cdf..4f897f848581 100644
> > --- a/drivers/media/pci/ttpci/budget-av.c
> > +++ b/drivers/media/pci/ttpci/budget-av.c
> > @@ -1463,7 +1463,8 @@ static int budget_av_attach(struct saa7146_dev *dev, struct saa7146_pci_extensio
> > budget_av->has_saa7113 = 1;
> > err = saa7146_vv_init(dev, &vv_data);
> > if (err != 0) {
> > - /* fixme: proper cleanup here */
> > + ttpci_budget_deinit(&budget_av->budget);
> > + kfree(budget_av);
> > ERR("cannot init vv subsystem\n");
> > return err;
> > }
> > @@ -1472,7 +1473,8 @@ static int budget_av_attach(struct saa7146_dev *dev, struct saa7146_pci_extensio
> > vv_data.vid_ops.vidioc_s_input = vidioc_s_input;
> >
> > if ((err = saa7146_register_device(&budget_av->vd, dev, "knc1", VFL_TYPE_VIDEO))) {
> > - /* fixme: proper cleanup here */
> > + ttpci_budget_deinit(&budget_av->budget);
> > + kfree(budget_av);
> I think this should go after the saa7146_vv_release() release below.
> Just in case saa7146_vv_release() relies on budget_av somewhere. And
> besides, it is good practice to release resources in the reverse order
> they were allocated.

Yes, I noticed the order of `kfree(budget_av)` and `ttpci_budget_deinit`,
but did forgot the `saa7146_vv_release`. I'll send a v2 version of this
patch soon.

Regards,
Zhipeng

>
> > ERR("cannot register capture v4l2 device\n");
> > saa7146_vv_release(dev);
> > return err;