diff mbox series

[04/17] reset: core: add get_device()/put_device on rcdev

Message ID 20240430083730.134918-5-herve.codina@bootlin.com (mailing list archive)
State New, archived
Headers show
Series Add support for the LAN966x PCI device using a DT overlay | expand

Commit Message

Herve Codina April 30, 2024, 8:37 a.m. UTC
From: Clément Léger <clement.leger@bootlin.com>

Since the rcdev structure is allocated by the reset controller drivers
themselves, they need to exists as long as there is a consumer. A call to
module_get() is already existing but that does not work when using
device-tree overlays. In order to guarantee that the underlying reset
controller device does not vanish while using it, add a get_device() call
when retrieving a reset control from a reset controller device and a
put_device() when releasing that control.

Signed-off-by: Clément Léger <clement.leger@bootlin.com>
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
---
 drivers/reset/core.c | 2 ++
 1 file changed, 2 insertions(+)

Comments

Philipp Zabel June 21, 2024, 10:16 a.m. UTC | #1
On Di, 2024-04-30 at 10:37 +0200, Herve Codina wrote:
> From: Clément Léger <clement.leger@bootlin.com>
> 
> Since the rcdev structure is allocated by the reset controller drivers
> themselves, they need to exists as long as there is a consumer. A call to
> module_get() is already existing but that does not work when using
> device-tree overlays. In order to guarantee that the underlying reset
> controller device does not vanish while using it, add a get_device() call
> when retrieving a reset control from a reset controller device and a
> put_device() when releasing that control.
> 
> Signed-off-by: Clément Léger <clement.leger@bootlin.com>
> Signed-off-by: Herve Codina <herve.codina@bootlin.com>
> ---
>  drivers/reset/core.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/reset/core.c b/drivers/reset/core.c
> index dba74e857be6..999c3c41cf21 100644
> --- a/drivers/reset/core.c
> +++ b/drivers/reset/core.c
> @@ -812,6 +812,7 @@ __reset_control_get_internal(struct reset_controller_dev *rcdev,
>  	kref_init(&rstc->refcnt);
>  	rstc->acquired = acquired;
>  	rstc->shared = shared;
> +	get_device(rcdev->dev);
> 

Looks good to me, but can we put this right after the try_module_get()
above ...
 
>  	return rstc;
>  }
> @@ -826,6 +827,7 @@ static void __reset_control_release(struct kref *kref)
>  	module_put(rstc->rcdev->owner);
>  
>  	list_del(&rstc->list);
> +	put_device(rstc->rcdev->dev);

... and this right before module_put()?

regards
Philipp
diff mbox series

Patch

diff --git a/drivers/reset/core.c b/drivers/reset/core.c
index dba74e857be6..999c3c41cf21 100644
--- a/drivers/reset/core.c
+++ b/drivers/reset/core.c
@@ -812,6 +812,7 @@  __reset_control_get_internal(struct reset_controller_dev *rcdev,
 	kref_init(&rstc->refcnt);
 	rstc->acquired = acquired;
 	rstc->shared = shared;
+	get_device(rcdev->dev);
 
 	return rstc;
 }
@@ -826,6 +827,7 @@  static void __reset_control_release(struct kref *kref)
 	module_put(rstc->rcdev->owner);
 
 	list_del(&rstc->list);
+	put_device(rstc->rcdev->dev);
 	kfree(rstc);
 }