| Message ID | 20231110152927.70601-35-u.kleine-koenig@pengutronix.de (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 77772addc1f110a3afc36a8883173f746361edb1 |
| Headers | show |
| Series | serial: Convert to platform remove callback returning void | expand |
On 10/11/2023 16:30, Uwe Kleine-König 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. Maybe let's trim the commit msg? Or convert all drivers at once if you want to drop long explanation... > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > --- Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Best regards, Krzysztof
On Sun, Nov 12, 2023 at 08:48:49AM +0100, Krzysztof Kozlowski wrote: > On 10/11/2023 16:30, Uwe Kleine-König 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. > > Maybe let's trim the commit msg? Or convert all drivers at once if you > want to drop long explanation... I like the verbosity of the commit log. This way you have a chance to understand its motivation by only looking at that single commit. IMHO doing one commit per driver is also right: It simplifies review, allows me to send patches to only the respective maintainers and so annoy less, the people reviewing the changes can send a tag for the driver change they look at only. But I'm not religious about that and if Greg wants to squash the patches together that's ok for me. Best regards Uwe
diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c index 3bd552841cd2..1b0c2b467a30 100644 --- a/drivers/tty/serial/samsung_tty.c +++ b/drivers/tty/serial/samsung_tty.c @@ -2054,7 +2054,7 @@ static int s3c24xx_serial_probe(struct platform_device *pdev) return 0; } -static int s3c24xx_serial_remove(struct platform_device *dev) +static void s3c24xx_serial_remove(struct platform_device *dev) { struct uart_port *port = s3c24xx_dev_to_port(&dev->dev); @@ -2063,8 +2063,6 @@ static int s3c24xx_serial_remove(struct platform_device *dev) } uart_unregister_driver(&s3c24xx_uart_drv); - - return 0; } /* UART power management code */ @@ -2611,7 +2609,7 @@ MODULE_DEVICE_TABLE(of, s3c24xx_uart_dt_match); static struct platform_driver samsung_serial_driver = { .probe = s3c24xx_serial_probe, - .remove = s3c24xx_serial_remove, + .remove_new = s3c24xx_serial_remove, .id_table = s3c24xx_serial_driver_ids, .driver = { .name = "samsung-uart",
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 <u.kleine-koenig@pengutronix.de> --- drivers/tty/serial/samsung_tty.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)