The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <[email protected]>
---
samples/qmi/qmi_sample_client.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/samples/qmi/qmi_sample_client.c b/samples/qmi/qmi_sample_client.c
index c045e3d24326..a42892523d3b 100644
--- a/samples/qmi/qmi_sample_client.c
+++ b/samples/qmi/qmi_sample_client.c
@@ -511,7 +511,7 @@ static int qmi_sample_probe(struct platform_device *pdev)
return ret;
}
-static int qmi_sample_remove(struct platform_device *pdev)
+static void qmi_sample_remove(struct platform_device *pdev)
{
struct qmi_sample *sample = platform_get_drvdata(pdev);
@@ -520,13 +520,11 @@ static int qmi_sample_remove(struct platform_device *pdev)
debugfs_remove(sample->de_dir);
qmi_handle_release(&sample->qmi);
-
- return 0;
}
static struct platform_driver qmi_sample_driver = {
.probe = qmi_sample_probe,
- .remove = qmi_sample_remove,
+ .remove_new = qmi_sample_remove,
.driver = {
.name = "qmi_sample_client",
},
base-commit: 8ffc8b1bbd505e27e2c8439d326b6059c906c9dd
--
2.43.0
Hello Bjorn,
On Fri, Mar 08, 2024 at 09:51:21AM +0100, Uwe Kleine-K?nig wrote:
> The .remove() callback for a platform driver returns an int which makes
> many driver authors wrongly assume it's possible to do error handling by
> returning an error code. However the value returned is ignored (apart
> from emitting a warning) and this typically results in resource leaks.
>
> To improve here there is a quest to make the remove callback return
> void. In the first step of this quest all drivers are converted to
> .remove_new(), which already returns void. Eventually after all drivers
> are converted, .remove_new() will be renamed to .remove().
>
> Trivially convert this driver from always returning zero in the remove
> callback to the void returning variant.
>
> Signed-off-by: Uwe Kleine-K?nig <[email protected]>
This patch didn't receive any feedback yet, I guess because I used an
old email address for Bjorn. I updated the address for this mail now.
Would you consider applying this patch; please tell me if I should
resend.
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-K?nig |
Industrial Linux Solutions | https://www.pengutronix.de/ |
Hello Bjorn,
On Fri, Apr 12, 2024 at 02:20:12PM +0200, Uwe Kleine-K?nig wrote:
> On Fri, Mar 08, 2024 at 09:51:21AM +0100, Uwe Kleine-K?nig wrote:
> > The .remove() callback for a platform driver returns an int which makes
> > many driver authors wrongly assume it's possible to do error handling by
> > returning an error code. However the value returned is ignored (apart
> > from emitting a warning) and this typically results in resource leaks.
> >
> > To improve here there is a quest to make the remove callback return
> > void. In the first step of this quest all drivers are converted to
> > .remove_new(), which already returns void. Eventually after all drivers
> > are converted, .remove_new() will be renamed to .remove().
> >
> > Trivially convert this driver from always returning zero in the remove
> > callback to the void returning variant.
> >
> > Signed-off-by: Uwe Kleine-K?nig <[email protected]>
>
> This patch didn't receive any feedback yet, I guess because I used an
> old email address for Bjorn. I updated the address for this mail now.
>
> Would you consider applying this patch; please tell me if I should
> resend.
I intend to change struct platform_driver after the upcoming merge
window. So if this patch doesn't go in, please accept me sending it
together with the actual change to struct platform_driver to Greg.
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-K?nig |
Industrial Linux Solutions | https://www.pengutronix.de/ |