2024-05-10 10:16:03

by Abdulrasaq Lawani

Subject: [PATCH] media: i2c: replacing of_node_put with __free(device_node)

Replaced instance of of_node_put with __free(device_node)
to protect against any memory leaks due to future changes
in control flow.

Suggested-by: Julia Lawall <[email protected]>
Signed-off-by: Abdulrasaq Lawani <[email protected]>
---
drivers/media/i2c/ov5647.c | 8 ++------
1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/drivers/media/i2c/ov5647.c b/drivers/media/i2c/ov5647.c
index 7e1ecdf2485f..d593dba092e3 100644
--- a/drivers/media/i2c/ov5647.c
+++ b/drivers/media/i2c/ov5647.c
@@ -1360,23 +1360,19 @@ static int ov5647_parse_dt(struct ov5647 *sensor, struct device_node *np)
struct v4l2_fwnode_endpoint bus_cfg = {
.bus_type = V4L2_MBUS_CSI2_DPHY,
};
- struct device_node *ep;
+ struct device_node *ep __free(device_node) = of_graph_get_endpoint_by_regs(np, 0, -1);
int ret;

- ep = of_graph_get_endpoint_by_regs(np, 0, -1);
if (!ep)
return -EINVAL;

ret = v4l2_fwnode_endpoint_parse(of_fwnode_handle(ep), &bus_cfg);
if (ret)
- goto out;
+ return ret;

sensor->clock_ncont = bus_cfg.bus.mipi_csi2.flags &
V4L2_MBUS_CSI2_NONCONTINUOUS_CLOCK;

-out:
- of_node_put(ep);
-
return ret;
}

--
2.34.1

2024-06-11 15:19:30

by Dave Stevenson

Subject: Re: [PATCH] media: i2c: replacing of_node_put with __free(device_node)

Hi

Thanks for the patch.

The subject should be "media: i2c: ov5647:" or just "media: ov5647" as
you're only touching the one driver.

On Fri, 10 May 2024 at 11:11, Abdulrasaq Lawani
<[email protected]> wrote:
>
> Replaced instance of of_node_put with __free(device_node)
> to protect against any memory leaks due to future changes
> in control flow.
>
> Suggested-by: Julia Lawall <[email protected]>
> Signed-off-by: Abdulrasaq Lawani <[email protected]>
> ---
> drivers/media/i2c/ov5647.c | 8 ++------
> 1 file changed, 2 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/media/i2c/ov5647.c b/drivers/media/i2c/ov5647.c
> index 7e1ecdf2485f..d593dba092e3 100644
> --- a/drivers/media/i2c/ov5647.c
> +++ b/drivers/media/i2c/ov5647.c
> @@ -1360,23 +1360,19 @@ static int ov5647_parse_dt(struct ov5647 *sensor, struct device_node *np)
> struct v4l2_fwnode_endpoint bus_cfg = {
> .bus_type = V4L2_MBUS_CSI2_DPHY,
> };
> - struct device_node *ep;
> + struct device_node *ep __free(device_node) = of_graph_get_endpoint_by_regs(np, 0, -1);

The media subsystem still requests a max line length of 80.
https://github.com/torvalds/linux/blob/master/Documentation/driver-api/media/maintainer-entry-profile.rst#coding-style-addendum
Break the line after the =

> int ret;
>
> - ep = of_graph_get_endpoint_by_regs(np, 0, -1);
> if (!ep)
> return -EINVAL;
>
> ret = v4l2_fwnode_endpoint_parse(of_fwnode_handle(ep), &bus_cfg);
> if (ret)
> - goto out;
> + return ret;
>
> sensor->clock_ncont = bus_cfg.bus.mipi_csi2.flags &
> V4L2_MBUS_CSI2_NONCONTINUOUS_CLOCK;
>
> -out:
> - of_node_put(ep);
> -
> return ret;

This could be "return 0;" as you've already returned if it was non-zero.

I know very little of this new mechanism, but it looks reasonable, and
Sakari has given it a basic blessing in a previous patch set.
With the above 3 comments addressed:
Acked-by: Dave Stevenson <[email protected]>

> }

>
> --
> 2.34.1
>