2023-01-11 12:06:16

by Greg Kroah-Hartman

[permalink] [raw]
Subject: [PATCH v2 03/16] platform/surface: aggregator: move to_ssam_device() to use container_of_const()

The driver core is changing to pass some pointers as const, so move
to_ssam_device() to use container_of_const() to handle this change.
to_ssam_device() now properly keeps the const-ness of the pointer passed
into it, while as before it could be lost.

Cc: Maximilian Luz <[email protected]>
Cc: [email protected]
Signed-off-by: Greg Kroah-Hartman <[email protected]>
---
include/linux/surface_aggregator/device.h | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/include/linux/surface_aggregator/device.h b/include/linux/surface_aggregator/device.h
index 46c45d1b6368..24151a0e2c96 100644
--- a/include/linux/surface_aggregator/device.h
+++ b/include/linux/surface_aggregator/device.h
@@ -229,10 +229,7 @@ static inline bool is_ssam_device(struct device *d)
* Return: Returns a pointer to the &struct ssam_device wrapping the given
* device @d.
*/
-static inline struct ssam_device *to_ssam_device(struct device *d)
-{
- return container_of(d, struct ssam_device, dev);
-}
+#define to_ssam_device(d) container_of_const(d, struct ssam_device, dev)

/**
* to_ssam_device_driver() - Casts the given device driver to a SSAM client
--
2.39.0


2023-01-11 12:21:57

by Hans de Goede

[permalink] [raw]
Subject: Re: [PATCH v2 03/16] platform/surface: aggregator: move to_ssam_device() to use container_of_const()

Hi,

On 1/11/23 12:30, Greg Kroah-Hartman wrote:
> The driver core is changing to pass some pointers as const, so move
> to_ssam_device() to use container_of_const() to handle this change.
> to_ssam_device() now properly keeps the const-ness of the pointer passed
> into it, while as before it could be lost.
>
> Cc: Maximilian Luz <[email protected]>
> Cc: [email protected]
> Signed-off-by: Greg Kroah-Hartman <[email protected]>
> ---
> include/linux/surface_aggregator/device.h | 5 +----
> 1 file changed, 1 insertion(+), 4 deletions(-)
>
> diff --git a/include/linux/surface_aggregator/device.h b/include/linux/surface_aggregator/device.h
> index 46c45d1b6368..24151a0e2c96 100644
> --- a/include/linux/surface_aggregator/device.h
> +++ b/include/linux/surface_aggregator/device.h
> @@ -229,10 +229,7 @@ static inline bool is_ssam_device(struct device *d)
> * Return: Returns a pointer to the &struct ssam_device wrapping the given
> * device @d.
> */
> -static inline struct ssam_device *to_ssam_device(struct device *d)
> -{
> - return container_of(d, struct ssam_device, dev);
> -}
> +#define to_ssam_device(d) container_of_const(d, struct ssam_device, dev)

There is a kernel-doc comment for this helper, I'm not sure if sphinx is going
to like changing this from a static inline to a #define ?

At a minimum I guess the generated docs are going to loose some type info
by switching to a define. Can we keep this as a static inline ?

Regards,

Hans




>
> /**
> * to_ssam_device_driver() - Casts the given device driver to a SSAM client

2023-01-11 12:26:32

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: [PATCH v2 03/16] platform/surface: aggregator: move to_ssam_device() to use container_of_const()

On Wed, Jan 11, 2023 at 01:02:01PM +0100, Maximilian Luz wrote:
> On 1/11/23 12:30, Greg Kroah-Hartman wrote:
> > The driver core is changing to pass some pointers as const, so move
> > to_ssam_device() to use container_of_const() to handle this change.
> > to_ssam_device() now properly keeps the const-ness of the pointer passed
> > into it, while as before it could be lost.
> >
> > Cc: Maximilian Luz <[email protected]>
> > Cc: [email protected]
> > Signed-off-by: Greg Kroah-Hartman <[email protected]>
> > ---
> > include/linux/surface_aggregator/device.h | 5 +----
> > 1 file changed, 1 insertion(+), 4 deletions(-)
> >
> > diff --git a/include/linux/surface_aggregator/device.h b/include/linux/surface_aggregator/device.h
> > index 46c45d1b6368..24151a0e2c96 100644
> > --- a/include/linux/surface_aggregator/device.h
> > +++ b/include/linux/surface_aggregator/device.h
> > @@ -229,10 +229,7 @@ static inline bool is_ssam_device(struct device *d)
> > * Return: Returns a pointer to the &struct ssam_device wrapping the given
> > * device @d.
> > */
> > -static inline struct ssam_device *to_ssam_device(struct device *d)
> > -{
> > - return container_of(d, struct ssam_device, dev);
> > -}
> > +#define to_ssam_device(d) container_of_const(d, struct ssam_device, dev)
> > /**
> > * to_ssam_device_driver() - Casts the given device driver to a SSAM client
>
> A small note in addition to the comment by Hans: I think it might be a
> good idea to convert the to_ssam_device_driver() function as well. I
> could also send this in as a follow-up patch.

Sure, that can be a follow-on patch as it's not needed here.

thanks,

greg k-h

2023-01-11 12:35:54

by Hans de Goede

[permalink] [raw]
Subject: Re: [PATCH v2 03/16] platform/surface: aggregator: move to_ssam_device() to use container_of_const()

Hi,

On 1/11/23 13:14, Greg Kroah-Hartman wrote:
> On Wed, Jan 11, 2023 at 12:54:54PM +0100, Hans de Goede wrote:
>> Hi,
>>
>> On 1/11/23 12:30, Greg Kroah-Hartman wrote:
>>> The driver core is changing to pass some pointers as const, so move
>>> to_ssam_device() to use container_of_const() to handle this change.
>>> to_ssam_device() now properly keeps the const-ness of the pointer passed
>>> into it, while as before it could be lost.
>>>
>>> Cc: Maximilian Luz <[email protected]>
>>> Cc: [email protected]
>>> Signed-off-by: Greg Kroah-Hartman <[email protected]>
>>> ---
>>> include/linux/surface_aggregator/device.h | 5 +----
>>> 1 file changed, 1 insertion(+), 4 deletions(-)
>>>
>>> diff --git a/include/linux/surface_aggregator/device.h b/include/linux/surface_aggregator/device.h
>>> index 46c45d1b6368..24151a0e2c96 100644
>>> --- a/include/linux/surface_aggregator/device.h
>>> +++ b/include/linux/surface_aggregator/device.h
>>> @@ -229,10 +229,7 @@ static inline bool is_ssam_device(struct device *d)
>>> * Return: Returns a pointer to the &struct ssam_device wrapping the given
>>> * device @d.
>>> */
>>> -static inline struct ssam_device *to_ssam_device(struct device *d)
>>> -{
>>> - return container_of(d, struct ssam_device, dev);
>>> -}
>>> +#define to_ssam_device(d) container_of_const(d, struct ssam_device, dev)
>>
>> There is a kernel-doc comment for this helper, I'm not sure if sphinx is going
>> to like changing this from a static inline to a #define ?
>>
>> At a minimum I guess the generated docs are going to loose some type info
>> by switching to a define. Can we keep this as a static inline ?
>
> No, container_of_const() will not work properly as an inline function as
> you don't know the return type until the preprocessor runs. I thought
> kerneldoc should still work just fine here as it does work for defines
> in other places.

Ok then this is fine as is. Thanks for clarifying why the switch to
a #define is necessary.

Regards,

Hans



2023-01-11 12:56:11

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: [PATCH v2 03/16] platform/surface: aggregator: move to_ssam_device() to use container_of_const()

On Wed, Jan 11, 2023 at 12:54:54PM +0100, Hans de Goede wrote:
> Hi,
>
> On 1/11/23 12:30, Greg Kroah-Hartman wrote:
> > The driver core is changing to pass some pointers as const, so move
> > to_ssam_device() to use container_of_const() to handle this change.
> > to_ssam_device() now properly keeps the const-ness of the pointer passed
> > into it, while as before it could be lost.
> >
> > Cc: Maximilian Luz <[email protected]>
> > Cc: [email protected]
> > Signed-off-by: Greg Kroah-Hartman <[email protected]>
> > ---
> > include/linux/surface_aggregator/device.h | 5 +----
> > 1 file changed, 1 insertion(+), 4 deletions(-)
> >
> > diff --git a/include/linux/surface_aggregator/device.h b/include/linux/surface_aggregator/device.h
> > index 46c45d1b6368..24151a0e2c96 100644
> > --- a/include/linux/surface_aggregator/device.h
> > +++ b/include/linux/surface_aggregator/device.h
> > @@ -229,10 +229,7 @@ static inline bool is_ssam_device(struct device *d)
> > * Return: Returns a pointer to the &struct ssam_device wrapping the given
> > * device @d.
> > */
> > -static inline struct ssam_device *to_ssam_device(struct device *d)
> > -{
> > - return container_of(d, struct ssam_device, dev);
> > -}
> > +#define to_ssam_device(d) container_of_const(d, struct ssam_device, dev)
>
> There is a kernel-doc comment for this helper, I'm not sure if sphinx is going
> to like changing this from a static inline to a #define ?
>
> At a minimum I guess the generated docs are going to loose some type info
> by switching to a define. Can we keep this as a static inline ?

No, container_of_const() will not work properly as an inline function as
you don't know the return type until the preprocessor runs. I thought
kerneldoc should still work just fine here as it does work for defines
in other places.

thanks,

greg k-h

2023-01-11 13:08:52

by Maximilian Luz

[permalink] [raw]
Subject: Re: [PATCH v2 03/16] platform/surface: aggregator: move to_ssam_device() to use container_of_const()

On 1/11/23 12:30, Greg Kroah-Hartman wrote:
> The driver core is changing to pass some pointers as const, so move
> to_ssam_device() to use container_of_const() to handle this change.
> to_ssam_device() now properly keeps the const-ness of the pointer passed
> into it, while as before it could be lost.
>
> Cc: Maximilian Luz <[email protected]>
> Cc: [email protected]
> Signed-off-by: Greg Kroah-Hartman <[email protected]>
> ---
> include/linux/surface_aggregator/device.h | 5 +----
> 1 file changed, 1 insertion(+), 4 deletions(-)
>
> diff --git a/include/linux/surface_aggregator/device.h b/include/linux/surface_aggregator/device.h
> index 46c45d1b6368..24151a0e2c96 100644
> --- a/include/linux/surface_aggregator/device.h
> +++ b/include/linux/surface_aggregator/device.h
> @@ -229,10 +229,7 @@ static inline bool is_ssam_device(struct device *d)
> * Return: Returns a pointer to the &struct ssam_device wrapping the given
> * device @d.
> */
> -static inline struct ssam_device *to_ssam_device(struct device *d)
> -{
> - return container_of(d, struct ssam_device, dev);
> -}
> +#define to_ssam_device(d) container_of_const(d, struct ssam_device, dev)
>
> /**
> * to_ssam_device_driver() - Casts the given device driver to a SSAM client

A small note in addition to the comment by Hans: I think it might be a
good idea to convert the to_ssam_device_driver() function as well. I
could also send this in as a follow-up patch.

Regards,
Max

2023-01-11 14:04:55

by Maximilian Luz

[permalink] [raw]
Subject: Re: [PATCH v2 03/16] platform/surface: aggregator: move to_ssam_device() to use container_of_const()

On 1/11/23 13:14, Greg Kroah-Hartman wrote:
> On Wed, Jan 11, 2023 at 01:02:01PM +0100, Maximilian Luz wrote:
>> On 1/11/23 12:30, Greg Kroah-Hartman wrote:
>>> The driver core is changing to pass some pointers as const, so move
>>> to_ssam_device() to use container_of_const() to handle this change.
>>> to_ssam_device() now properly keeps the const-ness of the pointer passed
>>> into it, while as before it could be lost.
>>>
>>> Cc: Maximilian Luz <[email protected]>
>>> Cc: [email protected]
>>> Signed-off-by: Greg Kroah-Hartman <[email protected]>
>>> ---
>>> include/linux/surface_aggregator/device.h | 5 +----
>>> 1 file changed, 1 insertion(+), 4 deletions(-)
>>>
>>> diff --git a/include/linux/surface_aggregator/device.h b/include/linux/surface_aggregator/device.h
>>> index 46c45d1b6368..24151a0e2c96 100644
>>> --- a/include/linux/surface_aggregator/device.h
>>> +++ b/include/linux/surface_aggregator/device.h
>>> @@ -229,10 +229,7 @@ static inline bool is_ssam_device(struct device *d)
>>> * Return: Returns a pointer to the &struct ssam_device wrapping the given
>>> * device @d.
>>> */
>>> -static inline struct ssam_device *to_ssam_device(struct device *d)
>>> -{
>>> - return container_of(d, struct ssam_device, dev);
>>> -}
>>> +#define to_ssam_device(d) container_of_const(d, struct ssam_device, dev)
>>> /**
>>> * to_ssam_device_driver() - Casts the given device driver to a SSAM client
>>
>> A small note in addition to the comment by Hans: I think it might be a
>> good idea to convert the to_ssam_device_driver() function as well. I
>> could also send this in as a follow-up patch.
>
> Sure, that can be a follow-on patch as it's not needed here.

Perfect, I'll do that.

As Hans' comment is now addressed, everything looks good to me.

Reviewed-by: Maximilian Luz <[email protected]>