[PATCH v3 0/2] Extend device_get_match_data() to struct bus_type

[PATCH v3 0/2] Extend device_get_match_data() to struct bus_type

[Biju Das]

This patch series extend device_get_match_data() to struct bus_type,
so that buses like I2C can get matched data.

v2->v3:
 * Added Rb tag from Andy for patch#1.
 * Extended to support i2c_of_match_device() as suggested by Andy.
 * Changed i2c_of_match_device_sysfs() as non-static function as it is
   needed for i2c_device_get_match_data().
 * Added a TODO comment to use i2c_verify_client() when it accepts const
   pointer.
 * Added multiple returns to make code path for device_get_match_data()
   faster in i2c_get_match_data().
RFC v1->v2:
 * Replaced "Signed-off-by"->"Suggested-by" tag for Dmitry.
 * Documented device_get_match_data().
 * Added multiple returns to make code path for generic fwnode-based
   lookup faster.
 * Fixed build warnings reported by kernel test robot <lkp@intel.com>
 * Added const qualifier to return type and parameter struct i2c_driver
   in i2c_get_match_data_helper().
 * Added const qualifier to struct i2c_driver in i2c_get_match_data()
 * Dropped driver variable from i2c_device_get_match_data()
 * Replaced to_i2c_client with logic for assigning verify_client as it
   returns non const pointer.

Biju Das (2):
  drivers: fwnode: Extend device_get_match_data() to struct bus_type
  i2c: Add i2c_device_get_match_data() callback

 drivers/base/property.c     | 21 +++++++++++++++-
 drivers/i2c/i2c-core-base.c | 49 +++++++++++++++++++++++++++++++------
 drivers/i2c/i2c-core-of.c   |  3 ++-
 include/linux/device/bus.h  |  3 +++
 include/linux/i2c.h         | 11 +++++++++
 5 files changed, 77 insertions(+), 10 deletions(-)

-- 
2.25.1

[PATCH v3 1/2] drivers: fwnode: Extend device_get_match_data() to struct bus_type

[Biju Das]

Extend device_get_match_data() to buses (for eg: I2C) by adding a
callback device_get_match_data() to struct bus_type() and call this method
as a fallback for generic fwnode based device_get_match_data().

Suggested-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
v2->v3:
 * Added Rb tag from Andy.
RFC v1-> v2:
 * Replaced "Signed-off-by"->"Suggested-by" tag for Dmitry.
 * Documented device_get_match_data().
 * Added multiple returns to make code path for generic fwnode-based
   lookup faster.
---
 drivers/base/property.c    | 21 ++++++++++++++++++++-
 include/linux/device/bus.h |  3 +++
 2 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/drivers/base/property.c b/drivers/base/property.c
index 8c40abed7852..e12af9ff0f51 100644
--- a/drivers/base/property.c
+++ b/drivers/base/property.c
@@ -1275,9 +1275,28 @@ int fwnode_graph_parse_endpoint(const struct fwnode_handle *fwnode,
 }
 EXPORT_SYMBOL(fwnode_graph_parse_endpoint);
 
+/**
+ * device_get_match_data - get match data from OF/ACPI/Bus match tables
+ * @dev: device to find the match data
+ *
+ * Find match data using generic fwnode-based lookup and if there is no
+ * match, call the bus->get_match_data() for finding match data.
+ *
+ * Return: a match data pointer or NULL if there is no match in the matching
+ * table.
+ */
 const void *device_get_match_data(const struct device *dev)
 {
-	return fwnode_call_ptr_op(dev_fwnode(dev), device_get_match_data, dev);
+	const void *data;
+
+	data = fwnode_call_ptr_op(dev_fwnode(dev), device_get_match_data, dev);
+	if (data)
+		return data;
+
+	if (dev->bus && dev->bus->get_match_data)
+		return dev->bus->get_match_data(dev);
+
+	return NULL;
 }
 EXPORT_SYMBOL_GPL(device_get_match_data);
 
diff --git a/include/linux/device/bus.h b/include/linux/device/bus.h
index ae10c4322754..2e15b0ae5384 100644
--- a/include/linux/device/bus.h
+++ b/include/linux/device/bus.h
@@ -60,6 +60,7 @@ struct fwnode_handle;
  *			this bus.
  * @dma_cleanup:	Called to cleanup DMA configuration on a device on
  *			this bus.
+ * @get_match_data:	Called to get match data on a device on this bus.
  * @pm:		Power management operations of this bus, callback the specific
  *		device driver's pm-ops.
  * @iommu_ops:  IOMMU specific operations for this bus, used to attach IOMMU
@@ -102,6 +103,8 @@ struct bus_type {
 	int (*dma_configure)(struct device *dev);
 	void (*dma_cleanup)(struct device *dev);
 
+	const void *(*get_match_data)(const struct device *dev);
+
 	const struct dev_pm_ops *pm;
 
 	const struct iommu_ops *iommu_ops;
-- 
2.25.1

Re: [PATCH v3 1/2] drivers: fwnode: Extend device_get_match_data() to struct bus_type

[Andy Shevchenko]

On Tue, Aug 01, 2023 at 06:03:17PM +0100, Biju Das wrote:
> Extend device_get_match_data() to buses (for eg: I2C) by adding a
> callback device_get_match_data() to struct bus_type() and call this method
> as a fallback for generic fwnode based device_get_match_data().

Because of anticipation of v4, see one additional comment below.

...

> +/**
> + * device_get_match_data - get match data from OF/ACPI/Bus match tables
> + * @dev: device to find the match data
> + *
> + * Find match data using generic fwnode-based lookup and if there is no
> + * match, call the bus->get_match_data() for finding match data.
> + *
> + * Return: a match data pointer or NULL if there is no match in the matching
> + * table.

Also add a note for the corner case.

"""
 *
 * Besides the fact that some drivers abuse the device ID driver_data type
 * and claim it to be integer, for the bus specific ID tables the driver_data
 * may be defined as kernel_ulong_t. For these tables 0 is a valid response,
 * but not for this function. It's recommended to covert those either to avoid
 * 0 or use a real pointer to the predefined driver data.
"""

> + */

-- 
With Best Regards,
Andy Shevchenko

RE: [PATCH v3 1/2] drivers: fwnode: Extend device_get_match_data() to struct bus_type

[Biju Das]

Hi Andy Shevchenko,

Thanks for the feedback.

> Subject: Re: [PATCH v3 1/2] drivers: fwnode: Extend
> device_get_match_data() to struct bus_type
> 
> On Tue, Aug 01, 2023 at 06:03:17PM +0100, Biju Das wrote:
> > Extend device_get_match_data() to buses (for eg: I2C) by adding a
> > callback device_get_match_data() to struct bus_type() and call this
> > method as a fallback for generic fwnode based device_get_match_data().
> 
> Because of anticipation of v4, see one additional comment below.

Ok.

> ...
> 
> > +/**
> > + * device_get_match_data - get match data from OF/ACPI/Bus match
> > +tables
> > + * @dev: device to find the match data
> > + *
> > + * Find match data using generic fwnode-based lookup and if there is
> > +no
> > + * match, call the bus->get_match_data() for finding match data.
> > + *
> > + * Return: a match data pointer or NULL if there is no match in the
> > +matching
> > + * table.
> 
> Also add a note for the corner case.
> 
> """
>  *
>  * Besides the fact that some drivers abuse the device ID driver_data
> type
>  * and claim it to be integer, for the bus specific ID tables the
> driver_data
>  * may be defined as kernel_ulong_t. For these tables 0 is a valid
> response,
>  * but not for this function. It's recommended to covert those either to

Will fix the typo covert->convert

Cheers,
Biju

> avoid
>  * 0 or use a real pointer to the predefined driver data.
> """
> 

Re: [PATCH v3 1/2] drivers: fwnode: Extend device_get_match_data() to struct bus_type

[Andy Shevchenko]

On Wed, Aug 02, 2023 at 06:32:18AM +0000, Biju Das wrote:
> > On Tue, Aug 01, 2023 at 06:03:17PM +0100, Biju Das wrote:

...

> >  *
> >  * Besides the fact that some drivers abuse the device ID driver_data
> > type
> >  * and claim it to be integer, for the bus specific ID tables the
> > driver_data
> >  * may be defined as kernel_ulong_t. For these tables 0 is a valid
> > response,
> >  * but not for this function. It's recommended to covert those either to
> 
> Will fix the typo covert->convert

Thank you!

> > avoid
> >  * 0 or use a real pointer to the predefined driver data.

-- 
With Best Regards,
Andy Shevchenko