On 2022/11/11 17:19, Laurent Pinchart wrote:
On Fri, Nov 11, 2022 at 04:58:09PM +0800, Yang Yingliang wrote:
Afer commit 1fa5ae857bb1 ("driver core: get rid of struct device's
bus_id string array"), the name of device is allocated dynamically,
so call kfree_const() to freed it in error case.
Fixes: 1fa5ae857bb1 ("driver core: get rid of struct device's bus_id string array")
Signed-off-by: Yang Yingliang <yangyingliang@xxxxxxxxxx>
---
drivers/media/v4l2-core/v4l2-dev.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/media/v4l2-core/v4l2-dev.c b/drivers/media/v4l2-core/v4l2-dev.c
index 397d553177fa..1fce1c29c60f 100644
--- a/drivers/media/v4l2-core/v4l2-dev.c
+++ b/drivers/media/v4l2-core/v4l2-dev.c
@@ -1036,6 +1036,9 @@ int __video_register_device(struct video_device *vdev,
ret = device_register(&vdev->dev);
if (ret < 0) {
pr_err("%s: device_register failed\n", __func__);
+ kfree_const(vdev->dev.kobj.name);
+ /* set it to null to avoid callers using a bad pointer */
+ vdev->dev.kobj.name = NULL;
This doesn't seem right to me. We shouldn't be poking into the kobj.
This should be handled in the driver core, or if that can't be done for
a reason specific to how the device is used in the V4L2 core, we need a
helper function from the driver core to perform the cleanup.
In technical, we should call put_device() here, but the release()
function has not been set
in this case, and v4l2_device_release() will release something that not
need be, so we can
not handle it in the driver core well, I think free the name here
directly is the best way to
fix it.
Thanks,
Yang
goto cleanup;
}
/* Register the release callback that will be called when the last
