diff mbox series

ndtest: Convert to platform remove callback returning void

Message ID c04bfc941a9f5d249b049572c1ae122fe551ee5d.1709886922.git.u.kleine-koenig@pengutronix.de (mailing list archive)
State Accepted
Commit 57456adef68db15ab81578d42848bae5b542999a
Delegated to: Ira Weiny
Headers show
Series ndtest: Convert to platform remove callback returning void | expand

Commit Message

Uwe Kleine-König March 8, 2024, 8:51 a.m. UTC
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>
---
 tools/testing/nvdimm/test/ndtest.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)


base-commit: 8ffc8b1bbd505e27e2c8439d326b6059c906c9dd

Comments

Dave Jiang March 29, 2024, 3:37 p.m. UTC | #1
On 3/8/24 1:51 AM, 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: Dave Jiang <dave.jiang@intel.com>
> ---
>  tools/testing/nvdimm/test/ndtest.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/tools/testing/nvdimm/test/ndtest.c b/tools/testing/nvdimm/test/ndtest.c
> index b8419f460368..2c6285aae852 100644
> --- a/tools/testing/nvdimm/test/ndtest.c
> +++ b/tools/testing/nvdimm/test/ndtest.c
> @@ -830,12 +830,11 @@ static int ndtest_bus_register(struct ndtest_priv *p)
>  	return 0;
>  }
>  
> -static int ndtest_remove(struct platform_device *pdev)
> +static void ndtest_remove(struct platform_device *pdev)
>  {
>  	struct ndtest_priv *p = to_ndtest_priv(&pdev->dev);
>  
>  	nvdimm_bus_unregister(p->bus);
> -	return 0;
>  }
>  
>  static int ndtest_probe(struct platform_device *pdev)
> @@ -882,7 +881,7 @@ static const struct platform_device_id ndtest_id[] = {
>  
>  static struct platform_driver ndtest_driver = {
>  	.probe = ndtest_probe,
> -	.remove = ndtest_remove,
> +	.remove_new = ndtest_remove,
>  	.driver = {
>  		.name = KBUILD_MODNAME,
>  	},
> 
> base-commit: 8ffc8b1bbd505e27e2c8439d326b6059c906c9dd
Dan Williams March 29, 2024, 6:45 p.m. UTC | #2
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>

Looks good,

Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Uwe Kleine-König April 12, 2024, 12:37 p.m. UTC | #3
On Fri, Mar 08, 2024 at 09:51:22AM +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>

This patch got two positive review feedback mails but isn't included in
next. This makes me wonder who feels responsible for picking it up.

Best regards
Uwe
diff mbox series

Patch

diff --git a/tools/testing/nvdimm/test/ndtest.c b/tools/testing/nvdimm/test/ndtest.c
index b8419f460368..2c6285aae852 100644
--- a/tools/testing/nvdimm/test/ndtest.c
+++ b/tools/testing/nvdimm/test/ndtest.c
@@ -830,12 +830,11 @@  static int ndtest_bus_register(struct ndtest_priv *p)
 	return 0;
 }
 
-static int ndtest_remove(struct platform_device *pdev)
+static void ndtest_remove(struct platform_device *pdev)
 {
 	struct ndtest_priv *p = to_ndtest_priv(&pdev->dev);
 
 	nvdimm_bus_unregister(p->bus);
-	return 0;
 }
 
 static int ndtest_probe(struct platform_device *pdev)
@@ -882,7 +881,7 @@  static const struct platform_device_id ndtest_id[] = {
 
 static struct platform_driver ndtest_driver = {
 	.probe = ndtest_probe,
-	.remove = ndtest_remove,
+	.remove_new = ndtest_remove,
 	.driver = {
 		.name = KBUILD_MODNAME,
 	},