[PATCH 00/13] irqchip: Convert to platform remove callback returning void

[PATCH 00/13] irqchip: Convert to platform remove callback returning void

[Uwe Kleine-König]

this series converts all drivers below drivers/irqchip to use
.remove_new(). See commit 5c5a7680e67b ("platform: Provide a remove
callback that returns no value") for an extended explanation and the
eventual goal. The TL;DR; is to make it harder for driver authors to
leak resources.

The drivers touched here are all fine though and don't return early in
.remove(). So all conversions in this series are trivial.

Best regards
Uwe

Uwe Kleine-König (13):
  irqchip/imgpdc: Convert to platform remove callback returning void
  irqchip/imx-intmux: Convert to platform remove callback returning void
  irqchip/imx-irqsteer: Convert to platform remove callback returning void
  irqchip/keystone: Convert to platform remove callback returning void
  irqchip/ls-scfg-msi: Convert to platform remove callback returning void
  irqchip/madera: Convert to platform remove callback returning void
  irqchip/mvebu-pic: Convert to platform remove callback returning void
  irqchip/pruss-intc: Convert to platform remove callback returning void
  irqchip/renesas-intc-irqpin: Convert to platform remove callback returning void
  irqchip/renesas-irqc: Convert to platform remove callback returning void
  irqchip/renesas-rza1: Convert to platform remove callback returning void
  irqchip/stm32-exti: Convert to platform remove callback returning void
  irqchip/ts4800: Convert to platform remove callback returning void

 drivers/irqchip/irq-imgpdc.c              | 5 ++---
 drivers/irqchip/irq-imx-intmux.c          | 6 ++----
 drivers/irqchip/irq-imx-irqsteer.c        | 6 ++----
 drivers/irqchip/irq-keystone.c            | 5 ++---
 drivers/irqchip/irq-ls-scfg-msi.c         | 6 ++----
 drivers/irqchip/irq-madera.c              | 6 ++----
 drivers/irqchip/irq-mvebu-pic.c           | 6 ++----
 drivers/irqchip/irq-pruss-intc.c          | 6 ++----
 drivers/irqchip/irq-renesas-intc-irqpin.c | 5 ++---
 drivers/irqchip/irq-renesas-irqc.c        | 5 ++---
 drivers/irqchip/irq-renesas-rza1.c        | 5 ++---
 drivers/irqchip/irq-stm32-exti.c          | 5 ++---
 drivers/irqchip/irq-ts4800.c              | 6 ++----
 13 files changed, 26 insertions(+), 46 deletions(-)

base-commit: 39676dfe52331dba909c617f213fdb21015c8d10
-- 
2.42.0

[PATCH 00/13] irqchip: Convert to platform remove callback returning void

[Uwe Kleine-König]

this series converts all drivers below drivers/irqchip to use
.remove_new(). See commit 5c5a7680e67b ("platform: Provide a remove
callback that returns no value") for an extended explanation and the
eventual goal. The TL;DR; is to make it harder for driver authors to
leak resources.

The drivers touched here are all fine though and don't return early in
.remove(). So all conversions in this series are trivial.

Best regards
Uwe

Uwe Kleine-König (13):
  irqchip/imgpdc: Convert to platform remove callback returning void
  irqchip/imx-intmux: Convert to platform remove callback returning void
  irqchip/imx-irqsteer: Convert to platform remove callback returning void
  irqchip/keystone: Convert to platform remove callback returning void
  irqchip/ls-scfg-msi: Convert to platform remove callback returning void
  irqchip/madera: Convert to platform remove callback returning void
  irqchip/mvebu-pic: Convert to platform remove callback returning void
  irqchip/pruss-intc: Convert to platform remove callback returning void
  irqchip/renesas-intc-irqpin: Convert to platform remove callback returning void
  irqchip/renesas-irqc: Convert to platform remove callback returning void
  irqchip/renesas-rza1: Convert to platform remove callback returning void
  irqchip/stm32-exti: Convert to platform remove callback returning void
  irqchip/ts4800: Convert to platform remove callback returning void

 drivers/irqchip/irq-imgpdc.c              | 5 ++---
 drivers/irqchip/irq-imx-intmux.c          | 6 ++----
 drivers/irqchip/irq-imx-irqsteer.c        | 6 ++----
 drivers/irqchip/irq-keystone.c            | 5 ++---
 drivers/irqchip/irq-ls-scfg-msi.c         | 6 ++----
 drivers/irqchip/irq-madera.c              | 6 ++----
 drivers/irqchip/irq-mvebu-pic.c           | 6 ++----
 drivers/irqchip/irq-pruss-intc.c          | 6 ++----
 drivers/irqchip/irq-renesas-intc-irqpin.c | 5 ++---
 drivers/irqchip/irq-renesas-irqc.c        | 5 ++---
 drivers/irqchip/irq-renesas-rza1.c        | 5 ++---
 drivers/irqchip/irq-stm32-exti.c          | 5 ++---
 drivers/irqchip/irq-ts4800.c              | 6 ++----
 13 files changed, 26 insertions(+), 46 deletions(-)

base-commit: 39676dfe52331dba909c617f213fdb21015c8d10
-- 
2.42.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH 01/13] irqchip/imgpdc: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-imgpdc.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-imgpdc.c b/drivers/irqchip/irq-imgpdc.c
index 5831be454673..be7451c52f3b 100644
--- a/drivers/irqchip/irq-imgpdc.c
+++ b/drivers/irqchip/irq-imgpdc.c
@@ -461,12 +461,11 @@ static int pdc_intc_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int pdc_intc_remove(struct platform_device *pdev)
+static void pdc_intc_remove(struct platform_device *pdev)
 {
 	struct pdc_intc_priv *priv = platform_get_drvdata(pdev);
 
 	irq_domain_remove(priv->domain);
-	return 0;
 }
 
 static const struct of_device_id pdc_intc_match[] = {
@@ -480,7 +479,7 @@ static struct platform_driver pdc_intc_driver = {
 		.of_match_table	= pdc_intc_match,
 	},
 	.probe = pdc_intc_probe,
-	.remove = pdc_intc_remove,
+	.remove_new = pdc_intc_remove,
 };
 
 static int __init pdc_intc_init(void)
-- 
2.42.0

[PATCH 02/13] irqchip/imx-intmux: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-imx-intmux.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-imx-intmux.c b/drivers/irqchip/irq-imx-intmux.c
index aa041e4dfee0..de292bde058f 100644
--- a/drivers/irqchip/irq-imx-intmux.c
+++ b/drivers/irqchip/irq-imx-intmux.c
@@ -282,7 +282,7 @@ static int imx_intmux_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int imx_intmux_remove(struct platform_device *pdev)
+static void imx_intmux_remove(struct platform_device *pdev)
 {
 	struct intmux_data *data = platform_get_drvdata(pdev);
 	int i;
@@ -298,8 +298,6 @@ static int imx_intmux_remove(struct platform_device *pdev)
 	}
 
 	pm_runtime_disable(&pdev->dev);
-
-	return 0;
 }
 
 #ifdef CONFIG_PM
@@ -359,6 +357,6 @@ static struct platform_driver imx_intmux_driver = {
 		.pm = &imx_intmux_pm_ops,
 	},
 	.probe = imx_intmux_probe,
-	.remove = imx_intmux_remove,
+	.remove_new = imx_intmux_remove,
 };
 builtin_platform_driver(imx_intmux_driver);
-- 
2.42.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH 02/13] irqchip/imx-intmux: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-imx-intmux.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-imx-intmux.c b/drivers/irqchip/irq-imx-intmux.c
index aa041e4dfee0..de292bde058f 100644
--- a/drivers/irqchip/irq-imx-intmux.c
+++ b/drivers/irqchip/irq-imx-intmux.c
@@ -282,7 +282,7 @@ static int imx_intmux_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int imx_intmux_remove(struct platform_device *pdev)
+static void imx_intmux_remove(struct platform_device *pdev)
 {
 	struct intmux_data *data = platform_get_drvdata(pdev);
 	int i;
@@ -298,8 +298,6 @@ static int imx_intmux_remove(struct platform_device *pdev)
 	}
 
 	pm_runtime_disable(&pdev->dev);
-
-	return 0;
 }
 
 #ifdef CONFIG_PM
@@ -359,6 +357,6 @@ static struct platform_driver imx_intmux_driver = {
 		.pm = &imx_intmux_pm_ops,
 	},
 	.probe = imx_intmux_probe,
-	.remove = imx_intmux_remove,
+	.remove_new = imx_intmux_remove,
 };
 builtin_platform_driver(imx_intmux_driver);
-- 
2.42.0

[PATCH 03/13] irqchip/imx-irqsteer: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-imx-irqsteer.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-imx-irqsteer.c b/drivers/irqchip/irq-imx-irqsteer.c
index bd9543314539..05f6b5e8370d 100644
--- a/drivers/irqchip/irq-imx-irqsteer.c
+++ b/drivers/irqchip/irq-imx-irqsteer.c
@@ -231,7 +231,7 @@ static int imx_irqsteer_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int imx_irqsteer_remove(struct platform_device *pdev)
+static void imx_irqsteer_remove(struct platform_device *pdev)
 {
 	struct irqsteer_data *irqsteer_data = platform_get_drvdata(pdev);
 	int i;
@@ -243,8 +243,6 @@ static int imx_irqsteer_remove(struct platform_device *pdev)
 	irq_domain_remove(irqsteer_data->domain);
 
 	clk_disable_unprepare(irqsteer_data->ipg_clk);
-
-	return 0;
 }
 
 #ifdef CONFIG_PM
@@ -312,6 +310,6 @@ static struct platform_driver imx_irqsteer_driver = {
 		.pm = &imx_irqsteer_pm_ops,
 	},
 	.probe = imx_irqsteer_probe,
-	.remove = imx_irqsteer_remove,
+	.remove_new = imx_irqsteer_remove,
 };
 builtin_platform_driver(imx_irqsteer_driver);
-- 
2.42.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH 03/13] irqchip/imx-irqsteer: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-imx-irqsteer.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-imx-irqsteer.c b/drivers/irqchip/irq-imx-irqsteer.c
index bd9543314539..05f6b5e8370d 100644
--- a/drivers/irqchip/irq-imx-irqsteer.c
+++ b/drivers/irqchip/irq-imx-irqsteer.c
@@ -231,7 +231,7 @@ static int imx_irqsteer_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int imx_irqsteer_remove(struct platform_device *pdev)
+static void imx_irqsteer_remove(struct platform_device *pdev)
 {
 	struct irqsteer_data *irqsteer_data = platform_get_drvdata(pdev);
 	int i;
@@ -243,8 +243,6 @@ static int imx_irqsteer_remove(struct platform_device *pdev)
 	irq_domain_remove(irqsteer_data->domain);
 
 	clk_disable_unprepare(irqsteer_data->ipg_clk);
-
-	return 0;
 }
 
 #ifdef CONFIG_PM
@@ -312,6 +310,6 @@ static struct platform_driver imx_irqsteer_driver = {
 		.pm = &imx_irqsteer_pm_ops,
 	},
 	.probe = imx_irqsteer_probe,
-	.remove = imx_irqsteer_remove,
+	.remove_new = imx_irqsteer_remove,
 };
 builtin_platform_driver(imx_irqsteer_driver);
-- 
2.42.0

[PATCH 04/13] irqchip/keystone: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-keystone.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-keystone.c b/drivers/irqchip/irq-keystone.c
index a36396db4b08..30f1979fa124 100644
--- a/drivers/irqchip/irq-keystone.c
+++ b/drivers/irqchip/irq-keystone.c
@@ -190,7 +190,7 @@ static int keystone_irq_probe(struct platform_device *pdev)
 	return 0;
 }
 
-static int keystone_irq_remove(struct platform_device *pdev)
+static void keystone_irq_remove(struct platform_device *pdev)
 {
 	struct keystone_irq_device *kirq = platform_get_drvdata(pdev);
 	int hwirq;
@@ -201,7 +201,6 @@ static int keystone_irq_remove(struct platform_device *pdev)
 		irq_dispose_mapping(irq_find_mapping(kirq->irqd, hwirq));
 
 	irq_domain_remove(kirq->irqd);
-	return 0;
 }
 
 static const struct of_device_id keystone_irq_dt_ids[] = {
@@ -212,7 +211,7 @@ MODULE_DEVICE_TABLE(of, keystone_irq_dt_ids);
 
 static struct platform_driver keystone_irq_device_driver = {
 	.probe		= keystone_irq_probe,
-	.remove		= keystone_irq_remove,
+	.remove_new	= keystone_irq_remove,
 	.driver		= {
 		.name	= "keystone_irq",
 		.of_match_table	= of_match_ptr(keystone_irq_dt_ids),
-- 
2.42.0

[PATCH 05/13] irqchip/ls-scfg-msi: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-ls-scfg-msi.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-ls-scfg-msi.c b/drivers/irqchip/irq-ls-scfg-msi.c
index 15cf80b46322..0b9b598b206f 100644
--- a/drivers/irqchip/irq-ls-scfg-msi.c
+++ b/drivers/irqchip/irq-ls-scfg-msi.c
@@ -398,7 +398,7 @@ static int ls_scfg_msi_probe(struct platform_device *pdev)
 	return 0;
 }
 
-static int ls_scfg_msi_remove(struct platform_device *pdev)
+static void ls_scfg_msi_remove(struct platform_device *pdev)
 {
 	struct ls_scfg_msi *msi_data = platform_get_drvdata(pdev);
 	int i;
@@ -410,8 +410,6 @@ static int ls_scfg_msi_remove(struct platform_device *pdev)
 	irq_domain_remove(msi_data->parent);
 
 	platform_set_drvdata(pdev, NULL);
-
-	return 0;
 }
 
 static struct platform_driver ls_scfg_msi_driver = {
@@ -420,7 +418,7 @@ static struct platform_driver ls_scfg_msi_driver = {
 		.of_match_table = ls_scfg_msi_id,
 	},
 	.probe = ls_scfg_msi_probe,
-	.remove = ls_scfg_msi_remove,
+	.remove_new = ls_scfg_msi_remove,
 };
 
 module_platform_driver(ls_scfg_msi_driver);
-- 
2.42.0

[PATCH 06/13] irqchip/madera: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-madera.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-madera.c b/drivers/irqchip/irq-madera.c
index 3eb1f8cdf674..b424c0a327f7 100644
--- a/drivers/irqchip/irq-madera.c
+++ b/drivers/irqchip/irq-madera.c
@@ -222,7 +222,7 @@ static int madera_irq_probe(struct platform_device *pdev)
 	return 0;
 }
 
-static int madera_irq_remove(struct platform_device *pdev)
+static void madera_irq_remove(struct platform_device *pdev)
 {
 	struct madera *madera = dev_get_drvdata(pdev->dev.parent);
 
@@ -232,13 +232,11 @@ static int madera_irq_remove(struct platform_device *pdev)
 	 */
 	madera->irq_dev = NULL;
 	regmap_del_irq_chip(madera->irq, madera->irq_data);
-
-	return 0;
 }
 
 static struct platform_driver madera_irq_driver = {
 	.probe	= &madera_irq_probe,
-	.remove = &madera_irq_remove,
+	.remove_new = madera_irq_remove,
 	.driver = {
 		.name	= "madera-irq",
 		.pm	= &madera_irq_pm_ops,
-- 
2.42.0

Re: [PATCH 06/13] irqchip/madera: Convert to platform remove callback returning void

[Richard Fitzgerald]

On 22/12/23 22:50, 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/irqchip/irq-madera.c | 6 ++----
>   1 file changed, 2 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/irqchip/irq-madera.c b/drivers/irqchip/irq-madera.c
> index 3eb1f8cdf674..b424c0a327f7 100644
> --- a/drivers/irqchip/irq-madera.c
> +++ b/drivers/irqchip/irq-madera.c
> @@ -222,7 +222,7 @@ static int madera_irq_probe(struct platform_device *pdev)
>   	return 0;
>   }
>   
> -static int madera_irq_remove(struct platform_device *pdev)
> +static void madera_irq_remove(struct platform_device *pdev)
>   {
>   	struct madera *madera = dev_get_drvdata(pdev->dev.parent);
>   
> @@ -232,13 +232,11 @@ static int madera_irq_remove(struct platform_device *pdev)
>   	 */
>   	madera->irq_dev = NULL;
>   	regmap_del_irq_chip(madera->irq, madera->irq_data);
> -
> -	return 0;
>   }
>   
>   static struct platform_driver madera_irq_driver = {
>   	.probe	= &madera_irq_probe,
> -	.remove = &madera_irq_remove,
> +	.remove_new = madera_irq_remove,
>   	.driver = {
>   		.name	= "madera-irq",
>   		.pm	= &madera_irq_pm_ops,

Reviewed-by: Richard Fitzgerald <rf@opensource.cirrus.com>

[PATCH 07/13] irqchip/mvebu-pic: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-mvebu-pic.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-mvebu-pic.c b/drivers/irqchip/irq-mvebu-pic.c
index ef3d3646ccc2..57e3f99b61f5 100644
--- a/drivers/irqchip/irq-mvebu-pic.c
+++ b/drivers/irqchip/irq-mvebu-pic.c
@@ -167,14 +167,12 @@ static int mvebu_pic_probe(struct platform_device *pdev)
 	return 0;
 }
 
-static int mvebu_pic_remove(struct platform_device *pdev)
+static void mvebu_pic_remove(struct platform_device *pdev)
 {
 	struct mvebu_pic *pic = platform_get_drvdata(pdev);
 
 	on_each_cpu(mvebu_pic_disable_percpu_irq, pic, 1);
 	irq_domain_remove(pic->domain);
-
-	return 0;
 }
 
 static const struct of_device_id mvebu_pic_of_match[] = {
@@ -185,7 +183,7 @@ MODULE_DEVICE_TABLE(of, mvebu_pic_of_match);
 
 static struct platform_driver mvebu_pic_driver = {
 	.probe  = mvebu_pic_probe,
-	.remove = mvebu_pic_remove,
+	.remove_new = mvebu_pic_remove,
 	.driver = {
 		.name = "mvebu-pic",
 		.of_match_table = mvebu_pic_of_match,
-- 
2.42.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH 07/13] irqchip/mvebu-pic: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-mvebu-pic.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-mvebu-pic.c b/drivers/irqchip/irq-mvebu-pic.c
index ef3d3646ccc2..57e3f99b61f5 100644
--- a/drivers/irqchip/irq-mvebu-pic.c
+++ b/drivers/irqchip/irq-mvebu-pic.c
@@ -167,14 +167,12 @@ static int mvebu_pic_probe(struct platform_device *pdev)
 	return 0;
 }
 
-static int mvebu_pic_remove(struct platform_device *pdev)
+static void mvebu_pic_remove(struct platform_device *pdev)
 {
 	struct mvebu_pic *pic = platform_get_drvdata(pdev);
 
 	on_each_cpu(mvebu_pic_disable_percpu_irq, pic, 1);
 	irq_domain_remove(pic->domain);
-
-	return 0;
 }
 
 static const struct of_device_id mvebu_pic_of_match[] = {
@@ -185,7 +183,7 @@ MODULE_DEVICE_TABLE(of, mvebu_pic_of_match);
 
 static struct platform_driver mvebu_pic_driver = {
 	.probe  = mvebu_pic_probe,
-	.remove = mvebu_pic_remove,
+	.remove_new = mvebu_pic_remove,
 	.driver = {
 		.name = "mvebu-pic",
 		.of_match_table = mvebu_pic_of_match,
-- 
2.42.0

Re: [PATCH 07/13] irqchip/mvebu-pic: Convert to platform remove callback returning void

[Gregory CLEMENT]

Uwe Kleine-König <u.kleine-koenig@pengutronix.de> writes:

> 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>

Acked-by: Gregory CLEMENT <gregory.clement@bootlin.com>

Thanks,

Gregory
> ---
>  drivers/irqchip/irq-mvebu-pic.c | 6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/irqchip/irq-mvebu-pic.c b/drivers/irqchip/irq-mvebu-pic.c
> index ef3d3646ccc2..57e3f99b61f5 100644
> --- a/drivers/irqchip/irq-mvebu-pic.c
> +++ b/drivers/irqchip/irq-mvebu-pic.c
> @@ -167,14 +167,12 @@ static int mvebu_pic_probe(struct platform_device *pdev)
>  	return 0;
>  }
>  
> -static int mvebu_pic_remove(struct platform_device *pdev)
> +static void mvebu_pic_remove(struct platform_device *pdev)
>  {
>  	struct mvebu_pic *pic = platform_get_drvdata(pdev);
>  
>  	on_each_cpu(mvebu_pic_disable_percpu_irq, pic, 1);
>  	irq_domain_remove(pic->domain);
> -
> -	return 0;
>  }
>  
>  static const struct of_device_id mvebu_pic_of_match[] = {
> @@ -185,7 +183,7 @@ MODULE_DEVICE_TABLE(of, mvebu_pic_of_match);
>  
>  static struct platform_driver mvebu_pic_driver = {
>  	.probe  = mvebu_pic_probe,
> -	.remove = mvebu_pic_remove,
> +	.remove_new = mvebu_pic_remove,
>  	.driver = {
>  		.name = "mvebu-pic",
>  		.of_match_table = mvebu_pic_of_match,
> -- 
> 2.42.0
>

-- 
Gregory Clement, Bootlin
Embedded Linux and Kernel engineering
http://bootlin.com

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 07/13] irqchip/mvebu-pic: Convert to platform remove callback returning void

[Gregory CLEMENT]

Uwe Kleine-König <u.kleine-koenig@pengutronix.de> writes:

> 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>

Acked-by: Gregory CLEMENT <gregory.clement@bootlin.com>

Thanks,

Gregory
> ---
>  drivers/irqchip/irq-mvebu-pic.c | 6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/irqchip/irq-mvebu-pic.c b/drivers/irqchip/irq-mvebu-pic.c
> index ef3d3646ccc2..57e3f99b61f5 100644
> --- a/drivers/irqchip/irq-mvebu-pic.c
> +++ b/drivers/irqchip/irq-mvebu-pic.c
> @@ -167,14 +167,12 @@ static int mvebu_pic_probe(struct platform_device *pdev)
>  	return 0;
>  }
>  
> -static int mvebu_pic_remove(struct platform_device *pdev)
> +static void mvebu_pic_remove(struct platform_device *pdev)
>  {
>  	struct mvebu_pic *pic = platform_get_drvdata(pdev);
>  
>  	on_each_cpu(mvebu_pic_disable_percpu_irq, pic, 1);
>  	irq_domain_remove(pic->domain);
> -
> -	return 0;
>  }
>  
>  static const struct of_device_id mvebu_pic_of_match[] = {
> @@ -185,7 +183,7 @@ MODULE_DEVICE_TABLE(of, mvebu_pic_of_match);
>  
>  static struct platform_driver mvebu_pic_driver = {
>  	.probe  = mvebu_pic_probe,
> -	.remove = mvebu_pic_remove,
> +	.remove_new = mvebu_pic_remove,
>  	.driver = {
>  		.name = "mvebu-pic",
>  		.of_match_table = mvebu_pic_of_match,
> -- 
> 2.42.0
>

-- 
Gregory Clement, Bootlin
Embedded Linux and Kernel engineering
http://bootlin.com

[PATCH 08/13] irqchip/pruss-intc: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-pruss-intc.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-pruss-intc.c b/drivers/irqchip/irq-pruss-intc.c
index 0f64ecb9b1f4..6daefa7c87bf 100644
--- a/drivers/irqchip/irq-pruss-intc.c
+++ b/drivers/irqchip/irq-pruss-intc.c
@@ -599,7 +599,7 @@ static int pruss_intc_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int pruss_intc_remove(struct platform_device *pdev)
+static void pruss_intc_remove(struct platform_device *pdev)
 {
 	struct pruss_intc *intc = platform_get_drvdata(pdev);
 	u8 max_system_events = intc->soc_config->num_system_events;
@@ -616,8 +616,6 @@ static int pruss_intc_remove(struct platform_device *pdev)
 		irq_dispose_mapping(irq_find_mapping(intc->domain, hwirq));
 
 	irq_domain_remove(intc->domain);
-
-	return 0;
 }
 
 static const struct pruss_intc_match_data pruss_intc_data = {
@@ -650,7 +648,7 @@ static struct platform_driver pruss_intc_driver = {
 		.suppress_bind_attrs = true,
 	},
 	.probe  = pruss_intc_probe,
-	.remove = pruss_intc_remove,
+	.remove_new = pruss_intc_remove,
 };
 module_platform_driver(pruss_intc_driver);
 
-- 
2.42.0

[PATCH 09/13] irqchip/renesas-intc-irqpin: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-renesas-intc-irqpin.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-renesas-intc-irqpin.c b/drivers/irqchip/irq-renesas-intc-irqpin.c
index fa19585f3dee..939293bc274e 100644
--- a/drivers/irqchip/irq-renesas-intc-irqpin.c
+++ b/drivers/irqchip/irq-renesas-intc-irqpin.c
@@ -561,14 +561,13 @@ static int intc_irqpin_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int intc_irqpin_remove(struct platform_device *pdev)
+static void intc_irqpin_remove(struct platform_device *pdev)
 {
 	struct intc_irqpin_priv *p = platform_get_drvdata(pdev);
 
 	irq_domain_remove(p->irq_domain);
 	pm_runtime_put(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
-	return 0;
 }
 
 static int __maybe_unused intc_irqpin_suspend(struct device *dev)
@@ -585,7 +584,7 @@ static SIMPLE_DEV_PM_OPS(intc_irqpin_pm_ops, intc_irqpin_suspend, NULL);
 
 static struct platform_driver intc_irqpin_device_driver = {
 	.probe		= intc_irqpin_probe,
-	.remove		= intc_irqpin_remove,
+	.remove_new	= intc_irqpin_remove,
 	.driver		= {
 		.name	= "renesas_intc_irqpin",
 		.of_match_table = intc_irqpin_dt_ids,
-- 
2.42.0

[PATCH 10/13] irqchip/renesas-irqc: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-renesas-irqc.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-renesas-irqc.c b/drivers/irqchip/irq-renesas-irqc.c
index 49b446b396f9..2bcf14d0f34b 100644
--- a/drivers/irqchip/irq-renesas-irqc.c
+++ b/drivers/irqchip/irq-renesas-irqc.c
@@ -218,14 +218,13 @@ static int irqc_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int irqc_remove(struct platform_device *pdev)
+static void irqc_remove(struct platform_device *pdev)
 {
 	struct irqc_priv *p = platform_get_drvdata(pdev);
 
 	irq_domain_remove(p->irq_domain);
 	pm_runtime_put(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
-	return 0;
 }
 
 static int __maybe_unused irqc_suspend(struct device *dev)
@@ -248,7 +247,7 @@ MODULE_DEVICE_TABLE(of, irqc_dt_ids);
 
 static struct platform_driver irqc_device_driver = {
 	.probe		= irqc_probe,
-	.remove		= irqc_remove,
+	.remove_new	= irqc_remove,
 	.driver		= {
 		.name	= "renesas_irqc",
 		.of_match_table	= irqc_dt_ids,
-- 
2.42.0

[PATCH 11/13] irqchip/renesas-rza1: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-renesas-rza1.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-renesas-rza1.c b/drivers/irqchip/irq-renesas-rza1.c
index e4c99c2e0373..f293ad13809f 100644
--- a/drivers/irqchip/irq-renesas-rza1.c
+++ b/drivers/irqchip/irq-renesas-rza1.c
@@ -244,12 +244,11 @@ static int rza1_irqc_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int rza1_irqc_remove(struct platform_device *pdev)
+static void rza1_irqc_remove(struct platform_device *pdev)
 {
 	struct rza1_irqc_priv *priv = platform_get_drvdata(pdev);
 
 	irq_domain_remove(priv->irq_domain);
-	return 0;
 }
 
 static const struct of_device_id rza1_irqc_dt_ids[] = {
@@ -260,7 +259,7 @@ MODULE_DEVICE_TABLE(of, rza1_irqc_dt_ids);
 
 static struct platform_driver rza1_irqc_device_driver = {
 	.probe		= rza1_irqc_probe,
-	.remove		= rza1_irqc_remove,
+	.remove_new	= rza1_irqc_remove,
 	.driver		= {
 		.name	= "renesas_rza1_irqc",
 		.of_match_table	= rza1_irqc_dt_ids,
-- 
2.42.0

[PATCH 12/13] irqchip/stm32-exti: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-stm32-exti.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-stm32-exti.c b/drivers/irqchip/irq-stm32-exti.c
index 971240e2e31b..c61a97caafc9 100644
--- a/drivers/irqchip/irq-stm32-exti.c
+++ b/drivers/irqchip/irq-stm32-exti.c
@@ -898,10 +898,9 @@ static void stm32_exti_remove_irq(void *data)
 	irq_domain_remove(domain);
 }
 
-static int stm32_exti_remove(struct platform_device *pdev)
+static void stm32_exti_remove(struct platform_device *pdev)
 {
 	stm32_exti_h_syscore_deinit();
-	return 0;
 }
 
 static int stm32_exti_probe(struct platform_device *pdev)
@@ -991,7 +990,7 @@ MODULE_DEVICE_TABLE(of, stm32_exti_ids);
 
 static struct platform_driver stm32_exti_driver = {
 	.probe		= stm32_exti_probe,
-	.remove		= stm32_exti_remove,
+	.remove_new	= stm32_exti_remove,
 	.driver		= {
 		.name	= "stm32_exti",
 		.of_match_table = stm32_exti_ids,
-- 
2.42.0


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH 12/13] irqchip/stm32-exti: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-stm32-exti.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/irqchip/irq-stm32-exti.c b/drivers/irqchip/irq-stm32-exti.c
index 971240e2e31b..c61a97caafc9 100644
--- a/drivers/irqchip/irq-stm32-exti.c
+++ b/drivers/irqchip/irq-stm32-exti.c
@@ -898,10 +898,9 @@ static void stm32_exti_remove_irq(void *data)
 	irq_domain_remove(domain);
 }
 
-static int stm32_exti_remove(struct platform_device *pdev)
+static void stm32_exti_remove(struct platform_device *pdev)
 {
 	stm32_exti_h_syscore_deinit();
-	return 0;
 }
 
 static int stm32_exti_probe(struct platform_device *pdev)
@@ -991,7 +990,7 @@ MODULE_DEVICE_TABLE(of, stm32_exti_ids);
 
 static struct platform_driver stm32_exti_driver = {
 	.probe		= stm32_exti_probe,
-	.remove		= stm32_exti_remove,
+	.remove_new	= stm32_exti_remove,
 	.driver		= {
 		.name	= "stm32_exti",
 		.of_match_table = stm32_exti_ids,
-- 
2.42.0

Re: [Linux-stm32] [PATCH 12/13] irqchip/stm32-exti: Convert to platform remove callback returning void

[Antonio Borneo]

On Fri, 2023-12-22 at 23:50 +0100, 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>

Reviewed-by: Antonio Borneo <antonio.borneo@foss.st.com>

Regards,
Antonio

> ---
>  drivers/irqchip/irq-stm32-exti.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/irqchip/irq-stm32-exti.c b/drivers/irqchip/irq-stm32-exti.c
> index 971240e2e31b..c61a97caafc9 100644
> --- a/drivers/irqchip/irq-stm32-exti.c
> +++ b/drivers/irqchip/irq-stm32-exti.c
> @@ -898,10 +898,9 @@ static void stm32_exti_remove_irq(void *data)
>         irq_domain_remove(domain);
>  }
>  
> -static int stm32_exti_remove(struct platform_device *pdev)
> +static void stm32_exti_remove(struct platform_device *pdev)
>  {
>         stm32_exti_h_syscore_deinit();
> -       return 0;
>  }
>  
>  static int stm32_exti_probe(struct platform_device *pdev)
> @@ -991,7 +990,7 @@ MODULE_DEVICE_TABLE(of, stm32_exti_ids);
>  
>  static struct platform_driver stm32_exti_driver = {
>         .probe          = stm32_exti_probe,
> -       .remove         = stm32_exti_remove,
> +       .remove_new     = stm32_exti_remove,
>         .driver         = {
>                 .name   = "stm32_exti",
>                 .of_match_table = stm32_exti_ids,


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [Linux-stm32] [PATCH 12/13] irqchip/stm32-exti: Convert to platform remove callback returning void

[Antonio Borneo]

On Fri, 2023-12-22 at 23:50 +0100, 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>

Reviewed-by: Antonio Borneo <antonio.borneo@foss.st.com>

Regards,
Antonio

> ---
>  drivers/irqchip/irq-stm32-exti.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/irqchip/irq-stm32-exti.c b/drivers/irqchip/irq-stm32-exti.c
> index 971240e2e31b..c61a97caafc9 100644
> --- a/drivers/irqchip/irq-stm32-exti.c
> +++ b/drivers/irqchip/irq-stm32-exti.c
> @@ -898,10 +898,9 @@ static void stm32_exti_remove_irq(void *data)
>         irq_domain_remove(domain);
>  }
>  
> -static int stm32_exti_remove(struct platform_device *pdev)
> +static void stm32_exti_remove(struct platform_device *pdev)
>  {
>         stm32_exti_h_syscore_deinit();
> -       return 0;
>  }
>  
>  static int stm32_exti_probe(struct platform_device *pdev)
> @@ -991,7 +990,7 @@ MODULE_DEVICE_TABLE(of, stm32_exti_ids);
>  
>  static struct platform_driver stm32_exti_driver = {
>         .probe          = stm32_exti_probe,
> -       .remove         = stm32_exti_remove,
> +       .remove_new     = stm32_exti_remove,
>         .driver         = {
>                 .name   = "stm32_exti",
>                 .of_match_table = stm32_exti_ids,

[PATCH 13/13] irqchip/ts4800: Convert to platform remove callback returning void

[Uwe Kleine-König]

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/irqchip/irq-ts4800.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/irqchip/irq-ts4800.c b/drivers/irqchip/irq-ts4800.c
index b2d61d4f6fe6..c69a819f099c 100644
--- a/drivers/irqchip/irq-ts4800.c
+++ b/drivers/irqchip/irq-ts4800.c
@@ -139,13 +139,11 @@ static int ts4800_ic_probe(struct platform_device *pdev)
 	return 0;
 }
 
-static int ts4800_ic_remove(struct platform_device *pdev)
+static void ts4800_ic_remove(struct platform_device *pdev)
 {
 	struct ts4800_irq_data *data = platform_get_drvdata(pdev);
 
 	irq_domain_remove(data->domain);
-
-	return 0;
 }
 
 static const struct of_device_id ts4800_ic_of_match[] = {
@@ -156,7 +154,7 @@ MODULE_DEVICE_TABLE(of, ts4800_ic_of_match);
 
 static struct platform_driver ts4800_ic_driver = {
 	.probe  = ts4800_ic_probe,
-	.remove = ts4800_ic_remove,
+	.remove_new = ts4800_ic_remove,
 	.driver = {
 		.name = "ts4800-irqc",
 		.of_match_table = ts4800_ic_of_match,
-- 
2.42.0

Re: [PATCH 00/13] irqchip: Convert to platform remove callback returning void

[Uwe Kleine-König]

[-- Attachment #1: Type: text/plain, Size: 911 bytes --]

Hello Thomas,

On Fri, Dec 22, 2023 at 11:50:31PM +0100, Uwe Kleine-König wrote:
> this series converts all drivers below drivers/irqchip to use
> .remove_new(). See commit 5c5a7680e67b ("platform: Provide a remove
> callback that returns no value") for an extended explanation and the
> eventual goal. The TL;DR; is to make it harder for driver authors to
> leak resources.
> 
> The drivers touched here are all fine though and don't return early in
> .remove(). So all conversions in this series are trivial.

I'm still waiting for this series to go in (or get review feedback). Is
this still on your radar? You're the right maintainer to take this
series, aren't you?

The series still applies to today's next.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH 00/13] irqchip: Convert to platform remove callback returning void

[Uwe Kleine-König]

[-- Attachment #1.1: Type: text/plain, Size: 911 bytes --]

Hello Thomas,

On Fri, Dec 22, 2023 at 11:50:31PM +0100, Uwe Kleine-König wrote:
> this series converts all drivers below drivers/irqchip to use
> .remove_new(). See commit 5c5a7680e67b ("platform: Provide a remove
> callback that returns no value") for an extended explanation and the
> eventual goal. The TL;DR; is to make it harder for driver authors to
> leak resources.
> 
> The drivers touched here are all fine though and don't return early in
> .remove(). So all conversions in this series are trivial.

I'm still waiting for this series to go in (or get review feedback). Is
this still on your radar? You're the right maintainer to take this
series, aren't you?

The series still applies to today's next.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[-- Attachment #2: Type: text/plain, Size: 176 bytes --]

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 00/13] irqchip: Convert to platform remove callback returning void

[Thomas Gleixner]

On Thu, Feb 15 2024 at 22:03, Uwe Kleine-König wrote:
> On Fri, Dec 22, 2023 at 11:50:31PM +0100, Uwe Kleine-König wrote:
>> this series converts all drivers below drivers/irqchip to use
>> .remove_new(). See commit 5c5a7680e67b ("platform: Provide a remove
>> callback that returns no value") for an extended explanation and the
>> eventual goal. The TL;DR; is to make it harder for driver authors to
>> leak resources.
>> 
>> The drivers touched here are all fine though and don't return early in
>> .remove(). So all conversions in this series are trivial.
>
> I'm still waiting for this series to go in (or get review feedback). Is
> this still on your radar? You're the right maintainer to take this
> series, aren't you?

I am and it fell through my christmas crack. I don't even try to catch
up with email after being almost 3 weeks AFK. For two decades I rely on
submitters to ping me after a couple of weeks or month in this case :)

Thanks,

        tglx

Re: [PATCH 00/13] irqchip: Convert to platform remove callback returning void

[Thomas Gleixner]

On Thu, Feb 15 2024 at 22:03, Uwe Kleine-König wrote:
> On Fri, Dec 22, 2023 at 11:50:31PM +0100, Uwe Kleine-König wrote:
>> this series converts all drivers below drivers/irqchip to use
>> .remove_new(). See commit 5c5a7680e67b ("platform: Provide a remove
>> callback that returns no value") for an extended explanation and the
>> eventual goal. The TL;DR; is to make it harder for driver authors to
>> leak resources.
>> 
>> The drivers touched here are all fine though and don't return early in
>> .remove(). So all conversions in this series are trivial.
>
> I'm still waiting for this series to go in (or get review feedback). Is
> this still on your radar? You're the right maintainer to take this
> series, aren't you?

I am and it fell through my christmas crack. I don't even try to catch
up with email after being almost 3 weeks AFK. For two decades I rely on
submitters to ping me after a couple of weeks or month in this case :)

Thanks,

        tglx


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel