On Thu, 2015-05-07 at 19:15 -0300, Fabio Estevam wrote:
> On Thu, May 7, 2015 at 6:51 PM, Srinivas Pandruvada
> <srinivas.pandruvada@xxxxxxxxxxxxxxx> wrote:
> > On Thu, 2015-05-07 at 11:09 -0300, Fabio Estevam wrote:
> >> 'channels' is allocated via kmemdup, so we need to free it after its usage.
> >>
> >> This way we fix the memory leak in both probe and remove paths.
> >>
> >> Also, as 'indio_dev' is allocated via devm_iio_device_alloc(), so there is
> >> no need to call kfree for it.
> > Where is the kfree for indio_dev..?
>
> It is kfree(indio_dev->channels) actually.
Then your commit message is wrong. devm_iio release will not free the
channels pointer.
>
> >>
> >> Reported-by: Alexey Khoroshilov <khoroshilov@xxxxxxxxx>
> >> Signed-off-by: Fabio Estevam <fabio.estevam@xxxxxxxxxxxxx>
> >> ---
> >> Changes since v1:
> >> - Always call kfree(channel) after its usage.
> >>
> >> drivers/iio/light/hid-sensor-prox.c | 14 ++++++--------
> >> 1 file changed, 6 insertions(+), 8 deletions(-)
> >>
> >> diff --git a/drivers/iio/light/hid-sensor-prox.c b/drivers/iio/light/hid-sensor-prox.c
> >> index 91ecc46..00ce0f6 100644
> >> --- a/drivers/iio/light/hid-sensor-prox.c
> >> +++ b/drivers/iio/light/hid-sensor-prox.c
> >> @@ -278,16 +278,17 @@ static int hid_prox_probe(struct platform_device *pdev)
> >> return -ENOMEM;
> >> }
> >>
> >> + indio_dev->channels = channels;
> >> +
> >> ret = prox_parse_report(pdev, hsdev, channels,
> >> HID_USAGE_SENSOR_PROX, prox_state);
> >> + kfree(channels);
> > You are freeing channels so indio_dev->channels will point to junk.
> > This will be passed to iio_device_register..
>
> No, I am freeing channels, not indio_dev->channels.
How can you do that? indio_dev->channels is pointer to the memory
allocated by channels..
>
> At this point in the code nobody uses channels, so it is safe to kfree it.
>
> >> if (ret) {
> >> dev_err(&pdev->dev, "failed to setup attributes\n");
> >> - goto error_free_dev_mem;
> >> + return ret;
> >> }
> >>
> >> - indio_dev->channels = channels;
> >> - indio_dev->num_channels =
> >> - ARRAY_SIZE(prox_channels);
> >> + indio_dev->num_channels = ARRAY_SIZE(prox_channels);
> >> indio_dev->dev.parent = &pdev->dev;
> >> indio_dev->info = &prox_info;
> >> indio_dev->name = name;
> >> @@ -297,7 +298,7 @@ static int hid_prox_probe(struct platform_device *pdev)
> >> NULL, NULL);
> >> if (ret) {
> >> dev_err(&pdev->dev, "failed to initialize trigger buffer\n");
> >> - goto error_free_dev_mem;
> >> + return ret;
> >> }
> >> atomic_set(&prox_state->common_attributes.data_ready, 0);
> >> ret = hid_sensor_setup_trigger(indio_dev, name,
> >> @@ -331,8 +332,6 @@ error_remove_trigger:
> >> hid_sensor_remove_trigger(&prox_state->common_attributes);
> >> error_unreg_buffer_funcs:
> >> iio_triggered_buffer_cleanup(indio_dev);
> >> -error_free_dev_mem:
> >> - kfree(indio_dev->channels);
> > This is required.
>
> Why? It is allocated with devm_ .
>
> >> return ret;
> >> }
> >>
> >> @@ -347,7 +346,6 @@ static int hid_prox_remove(struct platform_device *pdev)
> >> iio_device_unregister(indio_dev);
> >> hid_sensor_remove_trigger(&prox_state->common_attributes);
> >> iio_triggered_buffer_cleanup(indio_dev);
> >> - kfree(indio_dev->channels);
> >> This is required..
>
> Ditto.
--
To unsubscribe from this list: send the line "unsubscribe linux-iio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
