[PATCH 0/7] pcmcia: Convert to platform remove callback returning void

[PATCH 0/7] pcmcia: Convert to platform remove callback returning void

[Uwe Kleine-König]

Hello,

this series changes all platform drivers in drivers/pcmcia to use the
.remove_new() callback. See commit 5c5a7680e67b ("platform: Provide a
remove callback that returns no value") for an extended explanation and
the eventual goal.

All conversations are trivial, because all .remove() callbacks returned
zero unconditionally already.

There are no interdependencies between these patches, so they could be
picked up individually. However I'd expect them to go in all together.
It's unclrear to me though, who will pick them up. Dominik? Greg?

Best regards
Uwe

Uwe Kleine-König (7):
  pcmcia: bcm63xx: Convert to platform remove callback returning void
  pcmcia: db1xxx_ss: Convert to platform remove callback returning void
  pcmcia: electra_cf: Convert to platform remove callback returning void
  pcmcia: omap_cf: Convert to platform remove callback returning void
  pcmcia: pxa2xx: Convert to platform remove callback returning void
  pcmcia: sa1100: Convert to platform remove callback returning void
  pcmcia: xxs1500_ss: Convert to platform remove callback returning void

 drivers/pcmcia/bcm63xx_pcmcia.c | 5 ++---
 drivers/pcmcia/db1xxx_ss.c      | 6 ++----
 drivers/pcmcia/electra_cf.c     | 6 ++----
 drivers/pcmcia/omap_cf.c        | 5 ++---
 drivers/pcmcia/pxa2xx_base.c    | 6 ++----
 drivers/pcmcia/sa1100_generic.c | 8 +++-----
 drivers/pcmcia/xxs1500_ss.c     | 6 ++----
 7 files changed, 15 insertions(+), 27 deletions(-)

base-commit: 8e00ce02066e8f6f1ad5eab49a2ede7bf7a5ef64
-- 
2.42.0

[PATCH 1/7] pcmcia: bcm63xx: 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/pcmcia/bcm63xx_pcmcia.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/pcmcia/bcm63xx_pcmcia.c b/drivers/pcmcia/bcm63xx_pcmcia.c
index dd3c26099048..a5414441834a 100644
--- a/drivers/pcmcia/bcm63xx_pcmcia.c
+++ b/drivers/pcmcia/bcm63xx_pcmcia.c
@@ -437,7 +437,7 @@ static int bcm63xx_drv_pcmcia_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int bcm63xx_drv_pcmcia_remove(struct platform_device *pdev)
+static void bcm63xx_drv_pcmcia_remove(struct platform_device *pdev)
 {
 	struct bcm63xx_pcmcia_socket *skt;
 	struct resource *res;
@@ -449,12 +449,11 @@ static int bcm63xx_drv_pcmcia_remove(struct platform_device *pdev)
 	res = skt->reg_res;
 	release_mem_region(res->start, resource_size(res));
 	kfree(skt);
-	return 0;
 }
 
 struct platform_driver bcm63xx_pcmcia_driver = {
 	.probe	= bcm63xx_drv_pcmcia_probe,
-	.remove	= bcm63xx_drv_pcmcia_remove,
+	.remove_new = bcm63xx_drv_pcmcia_remove,
 	.driver	= {
 		.name	= "bcm63xx_pcmcia",
 		.owner  = THIS_MODULE,
-- 
2.42.0

[PATCH 2/7] pcmcia: db1xxx_ss: 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/pcmcia/db1xxx_ss.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/pcmcia/db1xxx_ss.c b/drivers/pcmcia/db1xxx_ss.c
index 87a33ecc2cf1..509713b9a502 100644
--- a/drivers/pcmcia/db1xxx_ss.c
+++ b/drivers/pcmcia/db1xxx_ss.c
@@ -577,7 +577,7 @@ static int db1x_pcmcia_socket_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int db1x_pcmcia_socket_remove(struct platform_device *pdev)
+static void db1x_pcmcia_socket_remove(struct platform_device *pdev)
 {
 	struct db1x_pcmcia_sock *sock = platform_get_drvdata(pdev);
 
@@ -585,8 +585,6 @@ static int db1x_pcmcia_socket_remove(struct platform_device *pdev)
 	pcmcia_unregister_socket(&sock->socket);
 	iounmap((void *)(sock->virt_io + (u32)mips_io_port_base));
 	kfree(sock);
-
-	return 0;
 }
 
 static struct platform_driver db1x_pcmcia_socket_driver = {
@@ -594,7 +592,7 @@ static struct platform_driver db1x_pcmcia_socket_driver = {
 		.name	= "db1xxx_pcmcia",
 	},
 	.probe		= db1x_pcmcia_socket_probe,
-	.remove		= db1x_pcmcia_socket_remove,
+	.remove_new	= db1x_pcmcia_socket_remove,
 };
 
 module_platform_driver(db1x_pcmcia_socket_driver);
-- 
2.42.0

[PATCH 3/7] pcmcia: electra_cf: 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/pcmcia/electra_cf.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/pcmcia/electra_cf.c b/drivers/pcmcia/electra_cf.c
index efc27bc15152..5ae826e54811 100644
--- a/drivers/pcmcia/electra_cf.c
+++ b/drivers/pcmcia/electra_cf.c
@@ -307,7 +307,7 @@ static int electra_cf_probe(struct platform_device *ofdev)
 
 }
 
-static int electra_cf_remove(struct platform_device *ofdev)
+static void electra_cf_remove(struct platform_device *ofdev)
 {
 	struct device *device = &ofdev->dev;
 	struct electra_cf_socket *cf;
@@ -326,8 +326,6 @@ static int electra_cf_remove(struct platform_device *ofdev)
 	release_region(cf->io_base, cf->io_size);
 
 	kfree(cf);
-
-	return 0;
 }
 
 static const struct of_device_id electra_cf_match[] = {
@@ -344,7 +342,7 @@ static struct platform_driver electra_cf_driver = {
 		.of_match_table = electra_cf_match,
 	},
 	.probe	  = electra_cf_probe,
-	.remove   = electra_cf_remove,
+	.remove_new = electra_cf_remove,
 };
 
 module_platform_driver(electra_cf_driver);
-- 
2.42.0

[PATCH 4/7] pcmcia: omap_cf: 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/pcmcia/omap_cf.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/pcmcia/omap_cf.c b/drivers/pcmcia/omap_cf.c
index e613818dc0bc..80137c7afe0d 100644
--- a/drivers/pcmcia/omap_cf.c
+++ b/drivers/pcmcia/omap_cf.c
@@ -290,7 +290,7 @@ static int __init omap_cf_probe(struct platform_device *pdev)
 	return status;
 }
 
-static int __exit omap_cf_remove(struct platform_device *pdev)
+static void __exit omap_cf_remove(struct platform_device *pdev)
 {
 	struct omap_cf_socket *cf = platform_get_drvdata(pdev);
 
@@ -300,14 +300,13 @@ static int __exit omap_cf_remove(struct platform_device *pdev)
 	release_mem_region(cf->phys_cf, SZ_8K);
 	free_irq(cf->irq, cf);
 	kfree(cf);
-	return 0;
 }
 
 static struct platform_driver omap_cf_driver = {
 	.driver = {
 		.name	= driver_name,
 	},
-	.remove		= __exit_p(omap_cf_remove),
+	.remove_new	= __exit_p(omap_cf_remove),
 };
 
 static int __init omap_cf_init(void)
-- 
2.42.0

[PATCH 5/7] pcmcia: pxa2xx: 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/pcmcia/pxa2xx_base.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/pcmcia/pxa2xx_base.c b/drivers/pcmcia/pxa2xx_base.c
index 5254028354f4..457fb81b497a 100644
--- a/drivers/pcmcia/pxa2xx_base.c
+++ b/drivers/pcmcia/pxa2xx_base.c
@@ -313,15 +313,13 @@ static int pxa2xx_drv_pcmcia_probe(struct platform_device *dev)
 	return ret;
 }
 
-static int pxa2xx_drv_pcmcia_remove(struct platform_device *dev)
+static void pxa2xx_drv_pcmcia_remove(struct platform_device *dev)
 {
 	struct skt_dev_info *sinfo = platform_get_drvdata(dev);
 	int i;
 
 	for (i = 0; i < sinfo->nskt; i++)
 		soc_pcmcia_remove_one(&sinfo->skt[i]);
-
-	return 0;
 }
 
 static int pxa2xx_drv_pcmcia_resume(struct device *dev)
@@ -338,7 +336,7 @@ static const struct dev_pm_ops pxa2xx_drv_pcmcia_pm_ops = {
 
 static struct platform_driver pxa2xx_pcmcia_driver = {
 	.probe		= pxa2xx_drv_pcmcia_probe,
-	.remove		= pxa2xx_drv_pcmcia_remove,
+	.remove_new	= pxa2xx_drv_pcmcia_remove,
 	.driver		= {
 		.name	= "pxa2xx-pcmcia",
 		.pm	= &pxa2xx_drv_pcmcia_pm_ops,
-- 
2.42.0

[PATCH 6/7] pcmcia: sa1100: 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/pcmcia/sa1100_generic.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/pcmcia/sa1100_generic.c b/drivers/pcmcia/sa1100_generic.c
index 89d4ba58c891..ccb219c38761 100644
--- a/drivers/pcmcia/sa1100_generic.c
+++ b/drivers/pcmcia/sa1100_generic.c
@@ -158,20 +158,18 @@ static int sa11x0_drv_pcmcia_probe(struct platform_device *pdev)
 	return sa11xx_drv_pcmcia_add_one(skt);
 }
 
-static int sa11x0_drv_pcmcia_remove(struct platform_device *dev)
+static void sa11x0_drv_pcmcia_remove(struct platform_device *dev)
 {
 	struct soc_pcmcia_socket *skt;
 
 	if (dev->id == -1) {
 		sa11x0_drv_pcmcia_legacy_remove(dev);
-		return 0;
+		return;
 	}
 
 	skt = platform_get_drvdata(dev);
 
 	soc_pcmcia_remove_one(skt);
-
-	return 0;
 }
 
 static struct platform_driver sa11x0_pcmcia_driver = {
@@ -179,7 +177,7 @@ static struct platform_driver sa11x0_pcmcia_driver = {
 		.name		= "sa11x0-pcmcia",
 	},
 	.probe		= sa11x0_drv_pcmcia_probe,
-	.remove		= sa11x0_drv_pcmcia_remove,
+	.remove_new	= sa11x0_drv_pcmcia_remove,
 };
 
 /* sa11x0_pcmcia_init()
-- 
2.42.0

[PATCH 7/7] pcmcia: xxs1500_ss: 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/pcmcia/xxs1500_ss.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/pcmcia/xxs1500_ss.c b/drivers/pcmcia/xxs1500_ss.c
index b11c7abb1dc0..2a93fbbd128d 100644
--- a/drivers/pcmcia/xxs1500_ss.c
+++ b/drivers/pcmcia/xxs1500_ss.c
@@ -301,7 +301,7 @@ static int xxs1500_pcmcia_probe(struct platform_device *pdev)
 	return ret;
 }
 
-static int xxs1500_pcmcia_remove(struct platform_device *pdev)
+static void xxs1500_pcmcia_remove(struct platform_device *pdev)
 {
 	struct xxs1500_pcmcia_sock *sock = platform_get_drvdata(pdev);
 
@@ -309,8 +309,6 @@ static int xxs1500_pcmcia_remove(struct platform_device *pdev)
 	free_irq(gpio_to_irq(GPIO_CDA), sock);
 	iounmap((void *)(sock->virt_io + (u32)mips_io_port_base));
 	kfree(sock);
-
-	return 0;
 }
 
 static struct platform_driver xxs1500_pcmcia_socket_driver = {
@@ -318,7 +316,7 @@ static struct platform_driver xxs1500_pcmcia_socket_driver = {
 		.name	= "xxs1500_pcmcia",
 	},
 	.probe		= xxs1500_pcmcia_probe,
-	.remove		= xxs1500_pcmcia_remove,
+	.remove_new	= xxs1500_pcmcia_remove,
 };
 
 module_platform_driver(xxs1500_pcmcia_socket_driver);
-- 
2.42.0

Re: [PATCH 0/7] pcmcia: Convert to platform remove callback returning void

[Dominik Brodowski]

Am Fri, Dec 08, 2023 at 05:08:05PM +0100 schrieb Uwe Kleine-König:
> Hello,
> 
> this series changes all platform drivers in drivers/pcmcia to use the
> .remove_new() callback. See commit 5c5a7680e67b ("platform: Provide a
> remove callback that returns no value") for an extended explanation and
> the eventual goal.
> 
> All conversations are trivial, because all .remove() callbacks returned
> zero unconditionally already.
> 
> There are no interdependencies between these patches, so they could be
> picked up individually. However I'd expect them to go in all together.
> It's unclrear to me though, who will pick them up. Dominik? Greg?

Both options are fine with me. In the latter case:

	Acked-by: Dominik Brodowski <linux@dominikbrodowski.net>

Best
	Dominik

Re: [PATCH 0/7] pcmcia: Convert to platform remove callback returning void

[Greg Kroah-Hartman]

On Fri, Dec 08, 2023 at 08:19:08PM +0100, Dominik Brodowski wrote:
> Am Fri, Dec 08, 2023 at 05:08:05PM +0100 schrieb Uwe Kleine-König:
> > Hello,
> > 
> > this series changes all platform drivers in drivers/pcmcia to use the
> > .remove_new() callback. See commit 5c5a7680e67b ("platform: Provide a
> > remove callback that returns no value") for an extended explanation and
> > the eventual goal.
> > 
> > All conversations are trivial, because all .remove() callbacks returned
> > zero unconditionally already.
> > 
> > There are no interdependencies between these patches, so they could be
> > picked up individually. However I'd expect them to go in all together.
> > It's unclrear to me though, who will pick them up. Dominik? Greg?
> 
> Both options are fine with me. In the latter case:
> 
> 	Acked-by: Dominik Brodowski <linux@dominikbrodowski.net>

I can take these, thanks!

greg k-h

Re: [PATCH 0/7] pcmcia: Convert to platform remove callback returning void

[Uwe Kleine-König]

[-- Attachment #1: Type: text/plain, Size: 1284 bytes --]

Hello Dominik,

On Fri, Dec 08, 2023 at 08:19:08PM +0100, Dominik Brodowski wrote:
> Am Fri, Dec 08, 2023 at 05:08:05PM +0100 schrieb Uwe Kleine-König:
> > Hello,
> > 
> > this series changes all platform drivers in drivers/pcmcia to use the
> > .remove_new() callback. See commit 5c5a7680e67b ("platform: Provide a
> > remove callback that returns no value") for an extended explanation and
> > the eventual goal.
> > 
> > All conversations are trivial, because all .remove() callbacks returned
> > zero unconditionally already.
> > 
> > There are no interdependencies between these patches, so they could be
> > picked up individually. However I'd expect them to go in all together.
> > It's unclrear to me though, who will pick them up. Dominik? Greg?
> 
> Both options are fine with me. In the latter case:
> 
> 	Acked-by: Dominik Brodowski <linux@dominikbrodowski.net>

Then I suggest you to take them. I only suggested Greg as a fallback
because I was unsure how active you are given the PCMCIA entry in
MAINTAINERS has status "Odd Fixes".

Best regards and a calm pre-Christmas period,
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 0/7] pcmcia: Convert to platform remove callback returning void

[Uwe Kleine-König]

[-- Attachment #1: Type: text/plain, Size: 594 bytes --]

Hello,

On Sat, Dec 09, 2023 at 10:28:36AM +0100, Greg Kroah-Hartman wrote:
> On Fri, Dec 08, 2023 at 08:19:08PM +0100, Dominik Brodowski wrote:
> > Both options are fine with me. In the latter case:
> > 
> > 	Acked-by: Dominik Brodowski <linux@dominikbrodowski.net>
> 
> I can take these, thanks!

I missed Greg's reply when I just suggested that Dominik should take
these. Either way is fine for me.

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 --]