1. Patchew
  2. linux
  3. [PATCH RFC 0/6] iio: mark scan_timestamp __private
  4. View patch

[PATCH RFC 6/6] iio: core: mark scan_timestamp as __private

Vasileios Amoiridis posted 6 patches 3 weeks, 5 days ago
[PATCH RFC 6/6] iio: core: mark scan_timestamp as __private
Posted by Vasileios Amoiridis 3 weeks, 5 days ago 
Since there are no more direct accesses to the indio_dev->scan_timestamp
value, it can be marked as __private and use the macro ACCESS_PRIVATE()
in order to access it. Like this, static checkers will be able to inform
in case someone tries to either write to the value, or read its value
directly.

Signed-off-by: Vasileios Amoiridis <vassilisamir@gmail.com>
---
 drivers/iio/industrialio-buffer.c | 2 +-
 include/linux/iio/iio.h           | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/iio/industrialio-buffer.c b/drivers/iio/industrialio-buffer.c
index 8104696cd475..c332741f3cf4 100644
--- a/drivers/iio/industrialio-buffer.c
+++ b/drivers/iio/industrialio-buffer.c
@@ -1137,7 +1137,7 @@ static int iio_enable_buffers(struct iio_dev *indio_dev,
 	int ret;
 
 	indio_dev->active_scan_mask = config->scan_mask;
-	indio_dev->scan_timestamp = config->scan_timestamp;
+	ACCESS_PRIVATE(indio_dev, scan_timestamp) = config->scan_timestamp;
 	indio_dev->scan_bytes = config->scan_bytes;
 	iio_dev_opaque->currentmode = config->mode;
 
diff --git a/include/linux/iio/iio.h b/include/linux/iio/iio.h
index 5661794d1127..669b4ef1280d 100644
--- a/include/linux/iio/iio.h
+++ b/include/linux/iio/iio.h
@@ -611,7 +611,7 @@ struct iio_dev {
 	const unsigned long		*available_scan_masks;
 	unsigned int			__private masklength;
 	const unsigned long		*active_scan_mask;
-	bool				scan_timestamp;
+	bool				__private scan_timestamp;
 	struct iio_trigger		*trig;
 	struct iio_poll_func		*pollfunc;
 	struct iio_poll_func		*pollfunc_event;
@@ -908,7 +908,7 @@ int iio_active_scan_mask_index(struct iio_dev *indio_dev);
  */
 static inline bool iio_is_soft_ts_enabled(const struct iio_dev *indio_dev)
 {
-	return indio_dev->scan_timestamp;
+	return ACCESS_PRIVATE(indio_dev, scan_timestamp);
 }
 
 ssize_t iio_format_value(char *buf, unsigned int type, int size, int *vals);
-- 
2.43.0
Re: [PATCH RFC 6/6] iio: core: mark scan_timestamp as __private
Posted by Jonathan Cameron 3 weeks, 4 days ago 
On Sat, 30 Nov 2024 01:27:10 +0100
Vasileios Amoiridis <vassilisamir@gmail.com> wrote:

> Since there are no more direct accesses to the indio_dev->scan_timestamp
> value, it can be marked as __private and use the macro ACCESS_PRIVATE()
> in order to access it. Like this, static checkers will be able to inform
> in case someone tries to either write to the value, or read its value
> directly.
> 
> Signed-off-by: Vasileios Amoiridis <vassilisamir@gmail.com>
> ---
>  drivers/iio/industrialio-buffer.c | 2 +-
>  include/linux/iio/iio.h           | 4 ++--
>  2 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/iio/industrialio-buffer.c b/drivers/iio/industrialio-buffer.c
> index 8104696cd475..c332741f3cf4 100644
> --- a/drivers/iio/industrialio-buffer.c
> +++ b/drivers/iio/industrialio-buffer.c
> @@ -1137,7 +1137,7 @@ static int iio_enable_buffers(struct iio_dev *indio_dev,
>  	int ret;
>  
>  	indio_dev->active_scan_mask = config->scan_mask;
> -	indio_dev->scan_timestamp = config->scan_timestamp;
> +	ACCESS_PRIVATE(indio_dev, scan_timestamp) = config->scan_timestamp;
>  	indio_dev->scan_bytes = config->scan_bytes;
>  	iio_dev_opaque->currentmode = config->mode;
>  
> diff --git a/include/linux/iio/iio.h b/include/linux/iio/iio.h
> index 5661794d1127..669b4ef1280d 100644
> --- a/include/linux/iio/iio.h
> +++ b/include/linux/iio/iio.h
> @@ -611,7 +611,7 @@ struct iio_dev {
>  	const unsigned long		*available_scan_masks;
>  	unsigned int			__private masklength;
>  	const unsigned long		*active_scan_mask;
> -	bool				scan_timestamp;
> +	bool				__private scan_timestamp;
>  	struct iio_trigger		*trig;
>  	struct iio_poll_func		*pollfunc;
>  	struct iio_poll_func		*pollfunc_event;
> @@ -908,7 +908,7 @@ int iio_active_scan_mask_index(struct iio_dev *indio_dev);
>   */
>  static inline bool iio_is_soft_ts_enabled(const struct iio_dev *indio_dev)
>  {
> -	return indio_dev->scan_timestamp;
> +	return ACCESS_PRIVATE(indio_dev, scan_timestamp);
If we only end up with one use of this (based on feedback on other drivers)
I'd tempted to deliberately not provide this convenience function and instead
just use ACCESS_PRIVATE() directly in iio_push_to_buffers_with_timestamp()

Nice work. Particularly by highlighting some 'odd corners' in drivers that
probably make no real sense to keep ;)

Jonathan


>  }
>  
>  ssize_t iio_format_value(char *buf, unsigned int type, int size, int *vals);
Re: [PATCH RFC 6/6] iio: core: mark scan_timestamp as __private
Posted by Vasileios Amoiridis 3 weeks, 1 day ago 
On Sat, Nov 30, 2024 at 02:19:54PM +0000, Jonathan Cameron wrote:
> On Sat, 30 Nov 2024 01:27:10 +0100
> Vasileios Amoiridis <vassilisamir@gmail.com> wrote:
> 
> > Since there are no more direct accesses to the indio_dev->scan_timestamp
> > value, it can be marked as __private and use the macro ACCESS_PRIVATE()
> > in order to access it. Like this, static checkers will be able to inform
> > in case someone tries to either write to the value, or read its value
> > directly.
> > 
> > Signed-off-by: Vasileios Amoiridis <vassilisamir@gmail.com>
> > ---
> >  drivers/iio/industrialio-buffer.c | 2 +-
> >  include/linux/iio/iio.h           | 4 ++--
> >  2 files changed, 3 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/iio/industrialio-buffer.c b/drivers/iio/industrialio-buffer.c
> > index 8104696cd475..c332741f3cf4 100644
> > --- a/drivers/iio/industrialio-buffer.c
> > +++ b/drivers/iio/industrialio-buffer.c
> > @@ -1137,7 +1137,7 @@ static int iio_enable_buffers(struct iio_dev *indio_dev,
> >  	int ret;
> >  
> >  	indio_dev->active_scan_mask = config->scan_mask;
> > -	indio_dev->scan_timestamp = config->scan_timestamp;
> > +	ACCESS_PRIVATE(indio_dev, scan_timestamp) = config->scan_timestamp;
> >  	indio_dev->scan_bytes = config->scan_bytes;
> >  	iio_dev_opaque->currentmode = config->mode;
> >  
> > diff --git a/include/linux/iio/iio.h b/include/linux/iio/iio.h
> > index 5661794d1127..669b4ef1280d 100644
> > --- a/include/linux/iio/iio.h
> > +++ b/include/linux/iio/iio.h
> > @@ -611,7 +611,7 @@ struct iio_dev {
> >  	const unsigned long		*available_scan_masks;
> >  	unsigned int			__private masklength;
> >  	const unsigned long		*active_scan_mask;
> > -	bool				scan_timestamp;
> > +	bool				__private scan_timestamp;
> >  	struct iio_trigger		*trig;
> >  	struct iio_poll_func		*pollfunc;
> >  	struct iio_poll_func		*pollfunc_event;
> > @@ -908,7 +908,7 @@ int iio_active_scan_mask_index(struct iio_dev *indio_dev);
> >   */
> >  static inline bool iio_is_soft_ts_enabled(const struct iio_dev *indio_dev)
> >  {
> > -	return indio_dev->scan_timestamp;
> > +	return ACCESS_PRIVATE(indio_dev, scan_timestamp);
> If we only end up with one use of this (based on feedback on other drivers)
> I'd tempted to deliberately not provide this convenience function and instead
> just use ACCESS_PRIVATE() directly in iio_push_to_buffers_with_timestamp()
> 
> Nice work. Particularly by highlighting some 'odd corners' in drivers that
> probably make no real sense to keep ;)
> 
> Jonathan
> 
> 
> >  }
> >  
> >  ssize_t iio_format_value(char *buf, unsigned int type, int size, int *vals);
>

Hi Jonathan,

Indeed, if it is only one case that this is being used, it wouldn't make
sense to provide an accessor. I wouldn't think of going directly to
touch the drivers without sending this RFC first, so it's good that you
like the solution of optimizing the drivers themselves. I might find
some time before the weekend to spin a v2 to discuss. Thanks for your
time, your comments are always of great help!

Cheers,
Vasilis

Patchew 2.1-devel-b67e4c2

© 2016 - 2024 Red Hat, Inc.