| Message ID | 20231110152927.70601-53-u.kleine-koenig@pengutronix.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | None | expand |
On 11/10/23 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. > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > --- > drivers/tty/serial/xilinx_uartps.c | 5 ++--- > 1 file changed, 2 insertions(+), 3 deletions(-) > > diff --git a/drivers/tty/serial/xilinx_uartps.c b/drivers/tty/serial/xilinx_uartps.c > index 66a45a634158..6decf2b13340 100644 > --- a/drivers/tty/serial/xilinx_uartps.c > +++ b/drivers/tty/serial/xilinx_uartps.c > @@ -1666,7 +1666,7 @@ static int cdns_uart_probe(struct platform_device *pdev) > * > * Return: 0 on success, negative errno otherwise this line should be also removed. > */ > -static int cdns_uart_remove(struct platform_device *pdev) > +static void cdns_uart_remove(struct platform_device *pdev) > { > struct uart_port *port = platform_get_drvdata(pdev); > struct cdns_uart *cdns_uart_data = port->private_data; > @@ -1692,12 +1692,11 @@ static int cdns_uart_remove(struct platform_device *pdev) > > if (!--instances) > uart_unregister_driver(cdns_uart_data->cdns_uart_driver); > - return 0; > } > > static struct platform_driver cdns_uart_platform_driver = { > .probe = cdns_uart_probe, > - .remove = cdns_uart_remove, > + .remove_new = cdns_uart_remove, > .driver = { > .name = CDNS_UART_NAME, > .of_match_table = cdns_uart_of_match, Thanks, Michal
On Mon, Nov 13, 2023 at 10:41:17AM +0100, Michal Simek wrote: > > > On 11/10/23 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. > > > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > > --- > > drivers/tty/serial/xilinx_uartps.c | 5 ++--- > > 1 file changed, 2 insertions(+), 3 deletions(-) > > > > diff --git a/drivers/tty/serial/xilinx_uartps.c b/drivers/tty/serial/xilinx_uartps.c > > index 66a45a634158..6decf2b13340 100644 > > --- a/drivers/tty/serial/xilinx_uartps.c > > +++ b/drivers/tty/serial/xilinx_uartps.c > > @@ -1666,7 +1666,7 @@ static int cdns_uart_probe(struct platform_device *pdev) > > * > > * Return: 0 on success, negative errno otherwise > > this line should be also removed. Good catch. @gregkh: What is the preferred way to address that? I can follow up with a patch on top of this one (and you squash it or not), I can send a patch that gets that does the conversion and drops the comment (and you skip this incomplete patch when applying this series) or I can resend the whole series in a week or so. Best regards Uwe
On Mon, Nov 13, 2023 at 11:03:42AM +0100, Uwe Kleine-König wrote: > On Mon, Nov 13, 2023 at 10:41:17AM +0100, Michal Simek wrote: > > > > > > On 11/10/23 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. > > > > > > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > > > --- > > > drivers/tty/serial/xilinx_uartps.c | 5 ++--- > > > 1 file changed, 2 insertions(+), 3 deletions(-) > > > > > > diff --git a/drivers/tty/serial/xilinx_uartps.c b/drivers/tty/serial/xilinx_uartps.c > > > index 66a45a634158..6decf2b13340 100644 > > > --- a/drivers/tty/serial/xilinx_uartps.c > > > +++ b/drivers/tty/serial/xilinx_uartps.c > > > @@ -1666,7 +1666,7 @@ static int cdns_uart_probe(struct platform_device *pdev) > > > * > > > * Return: 0 on success, negative errno otherwise > > > > this line should be also removed. > > Good catch. > > @gregkh: What is the preferred way to address that? I can follow up with > a patch on top of this one (and you squash it or not), I can send a > patch that gets that does the conversion and drops the comment (and you > skip this incomplete patch when applying this series) or I can resend > the whole series in a week or so. Follow-on patch is fine, I'll take this whole series in a week when I catch up from the merge-window-backlog as I'm at the Plumbers conference this week... thanks, greg k-h
diff --git a/drivers/tty/serial/xilinx_uartps.c b/drivers/tty/serial/xilinx_uartps.c index 66a45a634158..6decf2b13340 100644 --- a/drivers/tty/serial/xilinx_uartps.c +++ b/drivers/tty/serial/xilinx_uartps.c @@ -1666,7 +1666,7 @@ static int cdns_uart_probe(struct platform_device *pdev) * * Return: 0 on success, negative errno otherwise */ -static int cdns_uart_remove(struct platform_device *pdev) +static void cdns_uart_remove(struct platform_device *pdev) { struct uart_port *port = platform_get_drvdata(pdev); struct cdns_uart *cdns_uart_data = port->private_data; @@ -1692,12 +1692,11 @@ static int cdns_uart_remove(struct platform_device *pdev) if (!--instances) uart_unregister_driver(cdns_uart_data->cdns_uart_driver); - return 0; } static struct platform_driver cdns_uart_platform_driver = { .probe = cdns_uart_probe, - .remove = cdns_uart_remove, + .remove_new = cdns_uart_remove, .driver = { .name = CDNS_UART_NAME, .of_match_table = cdns_uart_of_match,
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/xilinx_uartps.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)