[RFC/PATCH] dt: Add of_device_compatible_match()

[RFC/PATCH] dt: Add of_device_compatible_match()

[Benjamin Herrenschmidt]

This provides an equivalent of of_fdt_match() for non-flat trees.

This is more practical than matching an array of of_device_id structs
when converting a bunch of existing users of of_fdt_match().

Signed-off-by: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org>
---

This is a pre-requisite for some work I'm doing to move the platform
identification in arch/powerpc until after we have unflattened the
device-tree. Rewriting all those lists as of_device_id's would be
significantly cumbersome.

Note: untested other than it compiles. I want an agreement on the
interface ASAP since the conversion of all the platforms is a burden
I'd like to avoid doing twice.

diff --git a/drivers/of/base.c b/drivers/of/base.c
index ebf84e3..429c594 100644
--- a/drivers/of/base.c
+++ b/drivers/of/base.c
@@ -493,6 +493,28 @@ int of_device_is_compatible(const struct device_node *device,
 }
 EXPORT_SYMBOL(of_device_is_compatible);
 
+/** Checks if the device is compatible with any of the entries in
+ *  a NULL terminated array of strings. Returns the best match
+ *  score or 0.
+ */
+int of_device_compatible_match(struct device_node *device,
+			       const char *const *compat)
+{
+	unsigned int tmp, score = 0;
+
+	if (!compat)
+		return 0;
+
+	while (*compat) {
+		tmp = of_device_is_compatible(device, *compat);
+		if (tmp && (score == 0 || (tmp < score)))
+			score = tmp;
+		compat++;
+	}
+
+	return score;
+}
+
 /**
  * of_machine_is_compatible - Test root of device tree for a given compatible value
  * @compat: compatible string to look for in root node's compatible property.
diff --git a/include/linux/of.h b/include/linux/of.h
index 74eb28c..33c184d 100644
--- a/include/linux/of.h
+++ b/include/linux/of.h
@@ -324,6 +324,8 @@ extern int of_property_read_string_helper(const struct device_node *np,
 					      const char **out_strs, size_t sz, int index);
 extern int of_device_is_compatible(const struct device_node *device,
 				   const char *);
+extern int of_device_compatible_match(struct device_node *device,
+				      const char *const *compat);
 extern bool of_device_is_available(const struct device_node *device);
 extern bool of_device_is_big_endian(const struct device_node *device);
 extern const void *of_get_property(const struct device_node *node,
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC/PATCH] dt: Add of_device_compatible_match()

[Michael Ellerman]

On Mon, 2016-06-27 at 08:47 +1000, Benjamin Herrenschmidt wrote:
> This provides an equivalent of of_fdt_match() for non-flat trees.
> 
> This is more practical than matching an array of of_device_id structs
> when converting a bunch of existing users of of_fdt_match().
> 
> Signed-off-by: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org>
> ---
> 
> This is a pre-requisite for some work I'm doing to move the platform
> identification in arch/powerpc until after we have unflattened the
> device-tree. Rewriting all those lists as of_device_id's would be
> significantly cumbersome.
> 
> Note: untested other than it compiles. I want an agreement on the
> interface ASAP since the conversion of all the platforms is a burden
> I'd like to avoid doing twice.


Any comments on this?

We need it as a pre-req for a powerpc series, so unless anyone yells I'll merge
it via there.

cheers

> 
> diff --git a/drivers/of/base.c b/drivers/of/base.c
> index ebf84e3..429c594 100644
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -493,6 +493,28 @@ int of_device_is_compatible(const struct device_node *device,
>  }
>  EXPORT_SYMBOL(of_device_is_compatible);
>  
> +/** Checks if the device is compatible with any of the entries in
> + *  a NULL terminated array of strings. Returns the best match
> + *  score or 0.
> + */
> +int of_device_compatible_match(struct device_node *device,
> +			       const char *const *compat)
> +{
> +	unsigned int tmp, score = 0;
> +
> +	if (!compat)
> +		return 0;
> +
> +	while (*compat) {
> +		tmp = of_device_is_compatible(device, *compat);
> +		if (tmp && (score == 0 || (tmp < score)))
> +			score = tmp;
> +		compat++;
> +	}
> +
> +	return score;
> +}
> +
>  /**
>   * of_machine_is_compatible - Test root of device tree for a given compatible value
>   * @compat: compatible string to look for in root node's compatible property.
> diff --git a/include/linux/of.h b/include/linux/of.h
> index 74eb28c..33c184d 100644
> --- a/include/linux/of.h
> +++ b/include/linux/of.h
> @@ -324,6 +324,8 @@ extern int of_property_read_string_helper(const struct device_node *np,
>  					      const char **out_strs, size_t sz, int index);
>  extern int of_device_is_compatible(const struct device_node *device,
>  				   const char *);
> +extern int of_device_compatible_match(struct device_node *device,
> +				      const char *const *compat);
>  extern bool of_device_is_available(const struct device_node *device);
>  extern bool of_device_is_big_endian(const struct device_node *device);
>  extern const void *of_get_property(const struct device_node *node,
> --
> To unsubscribe from this list: send the line "unsubscribe devicetree" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 


--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC/PATCH] dt: Add of_device_compatible_match()

[Arnd Bergmann]

On Thursday, July 7, 2016 4:29:22 PM CEST Michael Ellerman wrote:
> On Mon, 2016-06-27 at 08:47 +1000, Benjamin Herrenschmidt wrote:
> > This provides an equivalent of of_fdt_match() for non-flat trees.
> > 
> > This is more practical than matching an array of of_device_id structs
> > when converting a bunch of existing users of of_fdt_match().
> > 
> > Signed-off-by: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org>
> > ---
> > 
> > This is a pre-requisite for some work I'm doing to move the platform
> > identification in arch/powerpc until after we have unflattened the
> > device-tree. Rewriting all those lists as of_device_id's would be
> > significantly cumbersome.
> > 
> > Note: untested other than it compiles. I want an agreement on the
> > interface ASAP since the conversion of all the platforms is a burden
> > I'd like to avoid doing twice.
> 
> 
> Any comments on this?
> 
> We need it as a pre-req for a powerpc series, so unless anyone yells I'll merge
> it via there.
> 

I like the idea, we already have a number of drivers that would benefit
from this.

	Arnd

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC/PATCH] dt: Add of_device_compatible_match()

[Frank Rowand]

On 06/26/16 15:47, Benjamin Herrenschmidt wrote:
> This provides an equivalent of of_fdt_match() for non-flat trees.
> 
> This is more practical than matching an array of of_device_id structs
> when converting a bunch of existing users of of_fdt_match().
> 
> Signed-off-by: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org>
> ---
> 
> This is a pre-requisite for some work I'm doing to move the platform
> identification in arch/powerpc until after we have unflattened the
> device-tree. Rewriting all those lists as of_device_id's would be
> significantly cumbersome.
> 
> Note: untested other than it compiles. I want an agreement on the
> interface ASAP since the conversion of all the platforms is a burden
> I'd like to avoid doing twice.
> 
> diff --git a/drivers/of/base.c b/drivers/of/base.c
> index ebf84e3..429c594 100644
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -493,6 +493,28 @@ int of_device_is_compatible(const struct device_node *device,
>  }
>  EXPORT_SYMBOL(of_device_is_compatible);
>  
> +/** Checks if the device is compatible with any of the entries in
> + *  a NULL terminated array of strings. Returns the best match
> + *  score or 0.
> + */
> +int of_device_compatible_match(struct device_node *device,
> +			       const char *const *compat)
> +{
> +	unsigned int tmp, score = 0;
> +
> +	if (!compat)
> +		return 0;
> +
> +	while (*compat) {
> +		tmp = of_device_is_compatible(device, *compat);
> +		if (tmp && (score == 0 || (tmp < score)))
> +			score = tmp;
> +		compat++;
> +	}
> +
> +	return score;
> +}
> +

of_device_is_compatible() returns a higher value for a better match.

of_fdt_match() returns a lower value for a better match.

So checking for (tmp < score) will give you the opposite of what you want.

And if you are replacing of_fdt_match() with of_device_compatible_match(),
the caller will have the expectation that a smaller score is a better
match.


>  /**
>   * of_machine_is_compatible - Test root of device tree for a given compatible value
>   * @compat: compatible string to look for in root node's compatible property.
> diff --git a/include/linux/of.h b/include/linux/of.h
> index 74eb28c..33c184d 100644
> --- a/include/linux/of.h
> +++ b/include/linux/of.h
> @@ -324,6 +324,8 @@ extern int of_property_read_string_helper(const struct device_node *np,
>  					      const char **out_strs, size_t sz, int index);
>  extern int of_device_is_compatible(const struct device_node *device,
>  				   const char *);
> +extern int of_device_compatible_match(struct device_node *device,
> +				      const char *const *compat);
>  extern bool of_device_is_available(const struct device_node *device);
>  extern bool of_device_is_big_endian(const struct device_node *device);
>  extern const void *of_get_property(const struct device_node *node,
> 

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC/PATCH] dt: Add of_device_compatible_match()

[Frank Rowand]

On 07/06/16 23:29, Michael Ellerman wrote:
> On Mon, 2016-06-27 at 08:47 +1000, Benjamin Herrenschmidt wrote:
>> This provides an equivalent of of_fdt_match() for non-flat trees.
>>  
>> This is more practical than matching an array of of_device_id structs
>> when converting a bunch of existing users of of_fdt_match().
>>  
>> Signed-off-by: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org>
>> ---
>>  
>> This is a pre-requisite for some work I'm doing to move the platform
>> identification in arch/powerpc until after we have unflattened the
>> device-tree. Rewriting all those lists as of_device_id's would be
>> significantly cumbersome.
>>  
>> Note: untested other than it compiles. I want an agreement on the
>> interface ASAP since the conversion of all the platforms is a burden
>> I'd like to avoid doing twice.
> 
> 
> Any comments on this?
> 
> We need it as a pre-req for a powerpc series, so unless anyone yells I'll merge
> it via there.


Just replied to Benjamin.  The logic is not quite right.

Rob has been handling the device tree merges, so I'll let him comment on
whether it is ok for you to merge, once it is correct.

-Frank
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC/PATCH] dt: Add of_device_compatible_match()

[Benjamin Herrenschmidt]

On Thu, 2016-07-07 at 09:42 -0700, Frank Rowand wrote:
> 
> of_device_is_compatible() returns a higher value for a better match.
> 
> of_fdt_match() returns a lower value for a better match.

Ugh ... that smells of trainwreck...

> So checking for (tmp < score) will give you the opposite of what you want.

Yup. I'll fix it.

> And if you are replacing of_fdt_match() with of_device_compatible_match(),
> the caller will have the expectation that a smaller score is a better
> match.

The few callers I converted only really cared about exact matches.

Cheers,
Ben.

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html